Experimental Methods in
Wastewater Treatment
Experimental Methods in
Wastewater Treatment
Mark C. M. van Loosdrecht
Per H. Nielsen
Carlos M. Lopez-Vazquez
Damir Brdjanovic
Published by:
IWA Publishing
Alliance House
12 Caxton Street
London SW1H 0QS, UK
T: +44 (0) 20 7654 5500
F: +44 (0) 20 7654 5555
E: publications@iwap.co.uk
I: www.iwapublishing.com
First published 2016
© 2016 IWA Publishing
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the UK Copyright, Designs and Patents
Act (1998), no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior permission in writing of
the publisher, or, in the case of photographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK,
or in accordance with the terms of licenses issued by the appropriate reproduction rights organization outside the UK. Enquiries concerning reproduction
outside the terms stated here should be sent to IWA Publishing at the address printed above.
The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any
legal responsibility or liability for errors or omissions that may be made.
Disclaimer
The information provided and the opinions given in this publication are not necessarily those of IWA and IWA Publishing and should not be acted upon
without independent consideration and professional advice. IWA and IWA Publishing will not accept responsibility for any loss or damage suffered by any
person acting or refraining from acting upon any material contained in this publication.
British Library Cataloguing in Publication Data
A CIP catalogue record for this book is available from the British Library
Library of Congress Cataloguing in Publication Data
A catalogue record for this book is available from the Library of Congress
Cover design:
Graphic design:
Peter Stroo
Hans Emeis
ISBN: 9781780404745 (Hardback)
ISBN: 9781780404752 (eBook)
Preface
Wastewater treatment is a core technology for water
resources protection and reuse, as is clearly demonstrated
by the great success of its consequent implementation in
many countries worldwide. During the last decennia
scientific research has made vast progress in understanding
the complex and interdisciplinary aspects of the biological,
biochemical, chemical and mechanical processes involved.
It can be concluded that the global application of existing
knowledge and experience in wastewater treatment
technology will represent a cornerstone in future water
management, as expressed in the Strategic Development
Goals accepted by the UN in September 2015.
Only about one fifth of the wastewater produced
globally is currently being adequately treated. To achieve
the goal for sustainable water management by 2030 would
require extra wastewater treatment facilities for about
600,000 people each day. I am convinced that this book
will make its own significant contribution to meeting this
ambitious goal.
In the near future, most of the global population will
live in cities and in low and middle-income countries,
where most wastewater is not adequately treated. Probably
the most limiting factor in achieving the goals for
sustainable water management is the lack of qualified,
well-trained professionals, able to comprehend the
scientific research results and transfer them into practice. It
is therefore of prime importance to make currently
available scientific advances and proven experiences in
wastewater treatment technology applications easily
accessible worldwide. This was one of the drivers for the
development of this book, which represents an innovative
contribution to help overcome such a capacity development
challenge. The book is most definitely expected to
contribute to bridging the gaps between the science and
technology, and their practical applications.
The great collection of authors and reviewers
represents an interdisciplinary team of globally
acknowledged experts. The book will therefore make a
major contribution to establishing a common professional
language, enhancing global communication between
wastewater professionals. In addition, the authors have
linked the description of the scientific basis for wastewater
treatment processes with a video-based online course for
the training of students, researchers, engineers, laboratory
technicians and treatment plant operators, demonstrating
commonly accepted experimentation procedures and their
application for lab-, pilot-, and full-scale treatment plant
operation.
From the perspective of the IWA this book also has the
great potential to enhance the development of a new
generation of researchers and enable them to communicate
on a global scale and beyond their specific field of
expertise. Both aspects are urgently needed to develop
adapted solutions for specific local conditions and to make
them globally available for implementation.
There has been a trend for some time that scientific
research and practice have been growing apart from each
other. Part of the reason for this is the global
implementation of an academic assessment method that
primarily focuses on the impact of publications on the
progress in scientific research. Applied research results
with an impact on practice in water quality management
are not yet being sufficiently rewarded as their impact is
not always reflected by citations in scientific journals. This
book attempts to overcome this problem as it aims to
enhance the dialogue and co-operation between scientists
and practitioners. Scientists are encouraged to deal with the
practical problems with scientific methods, while the
practitioners are encouraged to understand the scientific
background of all the processes relevant for treatment plant
optimization.
While conventional wastewater treatment plant
operation was driven by effluent quality and cost
minimization, this book fully incorporates the paradigm
shift towards material and energy recovery from
wastewater. In this respect the book is also very relevant
for developed countries, as the new paradigm will heavily
influence the future development of wastewater
management worldwide.
As IWA president I want to congratulate the authors of
this book on their great achievement and also thank the Bill
& Melinda Gates Foundation and the Dutch government
for their financial support.
Prof. Dr. Helmut Kroiss
President International Water Association
Contributors
Carlos M. Lopez-Vazquez
Damir Brdjanovic
Eldon R. Rene
Elena Ficara
Elena Torfs
Eveline I.P. Volcke
George A. Ekama
Glen T. Daigger
Gürkan Sin
Henri Spanjers
Holger Daims
Ilse Y. Smets
Imre Takács
Ingmar Nopens
Jeppe L. Nielsen
Jiři Wanner
Juan A. Baeza
Kartik Chandran
Krist V. Gernaey
Laurens Welles
Mads Albertsen
Mari K.H. Winkler
Mark C.M. van Loosdrecht
Mathieu Spérandio
Morten S. Dueholm
Nancy G. Love
Per H. Nielsen
Peter A. Vanrolleghem
Piet N.L. Lens
Rasmus H. Kirkegaard
Robert J. Seviour
Sebastiaan C.F. Meijer
Sophie Balemans
Søren M. Karst
Sylvie Gillot
Tessa P.H. van den Brand
Tommaso Lotti
Yves Comeau
UNESCO-IHE Institute for Water Education, The Netherlands
UNESCO-IHE Institute for Water Education, The Netherlands
UNESCO-IHE Institute for Water Education, The Netherlands
Milan University of Technology, Italy
Université Laval, Canada
Ghent University, Belgium
University of Cape Town, South Africa
University of Michigan, United States of America
Technical University of Denmark, Denmark
Delft University of Technology, The Netherlands
University of Vienna, Austria
Catholic University of Leuven, Belgium
Dynamita, France
Ghent University, Belgium
Aalborg University, Denmark
University of Chemistry and Technology Prague, Czech Republic
Universitat Autònoma de Barcelona, Spain
Columbia University, United States of America
Technical University of Denmark, Denmark
UNESCO-IHE Institute for Water Education, The Netherlands
Aalborg University, Denmark
University of Washington, United States of America
Delft University of Technology, The Netherlands
Institut National des Sciences Appliquées de Toulouse, France
Aalborg University, Denmark
University of Michigan, United States of America
Aalborg University, Denmark
Université Laval, Canada
UNESCO-IHE Institute for Water Education, The Netherlands
Aalborg University, Denmark
La Trobe University, Australia
Yuniko BV, The Netherlands
Ghent University, Belgium
Aalborg University, Denmark
IRSTEA, France
KWR Watercycle Research Institute, The Netherlands
Milan University of Technology, Italy
École Polytechnique de Montréal, Canada
1. 2.
1. 2.
2.
2.
6.
4.
3.
6.
5.
3.
8.
6.
6.
6.
7.
7.
5.
4.
5.
2.
8.
6.
1. 2. 4.
3.
8.
2.
1. 7. 8.
3. 4. 6.
2.
8.
7.
5.
6.
8.
4.
2.
2.
2.
Chapter author
Chapter reviewer
About the editors
Prof. Dr. Mark C.M. van Loosdrecht
Mark C.M. van Loosdrecht is a well-renown scientist recognised
for his significant contributions to the study of reducing energy
consumption and the footprint of wastewater treatment plants
through his patented and award-winning technologies Sharon®,
Anammox® and Nereda®. His main work focuses on the use of
microbial cultures within the environmental process-engineering
field, with a special emphasis on nutrient removal, biofilm and
biofouling. Currently he is a full professor and Group Leader of
Environmental Biotechnology at TU Delft. A fellow of the Royal
Dutch Academy of Arts and Sciences (KNAW), the Netherlands
Academy of Technology and Innovation (AcTI) and the
International Water Association (IWA), Professor van Loosdrecht
has won numerous prestigious awards. His research interests
include granular sludge systems, microbial storage polymers,
wastewater treatment, gas treatment, soil treatment, microbial
conversion of inorganic compounds, production of chemicals from
waste, and modelling. Apart from his other achievements, he has
published over 500 papers, supervised 65 PhD students so far and
is an honorary professor at the University of Queensland. He is
also currently the Editor-in-Chief for Water Research and Advisor
to IWA Publishing.
Prof. Dr. Per Halkjær Nielsen
Per H. Nielsen is a full professor at the Department of Chemistry
and Bioscience at Aalborg University, Denmark where he heads
the multidisciplinary Centre for Microbial Communities. He is also
a visiting scientist at the Singapore Centre on Environmental Life
Sciences
Engineering,
Nanyang
Technological
University, Singapore. Prof. Nielsen’s research group has been
active in environmental biotechnology for over 25 years, focusing
on the microbial ecology of biological wastewater treatment,
bioenergy production, bioremediation, biofilms, infection of
implants and the development of system microbiology approaches
based on new sequencing technologies. He chaired the IWA
specialist group Microbial Ecology and Water Engineering for
eight years (2005-2013) and is Chair of the IWA BioCluster. He is
a Fellow of the Danish Academy of Technical Sciences (ATV) and
the International Water Association (IWA) and has received
several prestigious awards. He has published more than 230 peerreviewed publications and supervised 25 PhD students. His main
research interest is microbial ecology in water engineering,
particularly related to wastewater treatment where he has
developed and applied several novel methods to study uncultured
microorganisms, e.g. by using next-generation sequencing
technologies. He is the initiator and responsible for the MiDAS
field guide open resource for wastewater microbiology.
Dr. Carlos M. Lopez-Vazquez
Carlos M. Lopez-Vazquez is Associate Professor in Wastewater
Treatment Technology at UNESCO-IHE Institute for Water
Education. In 2009 he received his doctoral degree on
Environmental Biotechnology (cum laude) from Delft University
of Technology and UNESCO-IHE Institute for Water Education.
During his professional career, he has taken part in different
advisory and consultancy projects for both public and private
sectors concerning municipal and industrial wastewater treatment
systems. After working for a couple of years in the Water R&D
Department of Nalco Europe on industrial water and wastewater
treatment applications, he re-joined UNESCO-IHE’s Sanitary
Engineering Chair Group in 2009. Since then, he has been
involved in education, capacity building and research projects
guiding dozens of MSc and several PhD students. By applying
mathematical modelling as an essential tool, he has a special focus
on the development and transfer of innovative and cost-effective
wastewater treatment technologies to developing countries,
countries in transition and industrial applications.
Prof. Dr. Damir Brdjanovic
Damir Brdjanovic is Professor of Sanitary Engineering at
UNESCO-IHE and Endowed Professor at Delft University of
Technology in the Environmental Biotechnology Group. Areas of
his expertise include pro-poor and emergency sanitation, faecal
sludge management, urban drainage, and wastewater treatment. He
is a pioneer in the practical application of models in wastewater
treatment practice in developing countries. He invented the Shit
Killer® device for excreta management in emergencies, the awardwinning eSOS® Smart Toilet and associated software eSOS
View®, with funding by the Bill & Melinda Gates Foundation
(BMGF). He has initiated the development and implementation of
innovative didactic approaches and novel educational products
(including e-learning) at UNESCO-IHE. In 2015, together with the
BMGF, he founded the Global Faecal Sludge Management elearning Alliance. Currently his chair group consists of ten staff
members, three post-doctoral fellows and 22 PhD students. In
addition, in excess of 100 MSc students have graduated under his
supervision so far. Prof. Brdjanovic has a sound publication
record, is co-initiator of the IWA Journal of Water, Sanitation and
Hygiene for Development, and is the initiator, author and editor of
five books in the wastewater treatment and sanitation field. In
2015 he became an International Water Association Fellow.
About the book and online course
Over the past twenty years, the knowledge and
understanding of wastewater treatment has advanced
extensively and moved away from empirically-based
approaches to a fundamentally-based first-principles
approach embracing chemistry, microbiology, and physical
and bioprocess engineering, often involving experimental
laboratory work and techniques. Many of these
experimental methods and techniques have matured to the
degree that they have been accepted as reliable tools in
wastewater treatment research and practice. For sector
professionals, especially the new generation of young
scientists and engineers entering the wastewater treatment
profession, the quantity, complexity and diversity of these
new developments can be overwhelming, particularly in
developing countries where access to advanced level
laboratory courses in wastewater treatment is not readily
available. In addition, information on innovative
experimental methods is scattered across scientific
literature and only partially available in the form of
textbooks or guidelines. This book seeks to address these
deficiencies. It assembles and integrates the innovative
experimental methods developed by research groups and
practitioners around the world and broadly applied in
wastewater treatment research and practice.
Experimental Methods in Wastewater Treatment book
forms part of the internet-based curriculum in sanitary
engineering at UNESCO-IHE and, as such, may also be
used together with video recordings of methods and
approaches performed and narrated by the authors,
including guidelines on best experimental practices. The
book is written for undergraduate and postgraduate
students, researchers, laboratory staff, plant operators,
consultants, and other sector professionals.
The idea of making this book and the online learning
course was conceived in 2009 when UNESCO-IHE agreed
to utilize some of the programmatic funds provided by the
Dutch Ministry of Foreign Affairs to develop innovative
learning methods and products. However it took until 2011
to acquire the additional funds from the Bill & Melinda
Gates Foundation (BMGF) that enabled the original idea to
be fully executed. The conceptual framework for the book,
and the online course that it is part of, was agreed upon in
Montreal during the IWA World Water Congress and
Exhibition in September 2010 and further detailed during
the IWA event in Essen, Activated Sludge – 100 Years and
Counting. The latter was the occasion when the concept
was introduced of also having established reviewers in the
field to provide critical feedback on the manuscripts and
improve the quality of the final product, in addition to the
esteemed groups of experts writing the chapters of the
book. Besides providing chapters in the book, authors were
requested to prepare presentation slides, tutorial exercises
and to deliver scenarios and narration for video-recorded
lectures and execution of experimental procedures at
UNESCO-IHE and partner laboratories. These materials
have been compiled into a digital package available to
those registered for the online course. IWA Publishing has
agreed to publish the book and market both the book and
online learning course. It has also been agreed that the
book and online course digital materials are available free
of charge. The online course is delivered once or twice a
year depending on the demand (please consult the
UNESCO-IHE website for further information on how to
embark on the course or download the course materials).
The book is also used for teaching as part of a lecture series
in the Sanitary Engineering specialization of the UNESCOIHE’s Master’s Program in Urban Water and Sanitation. It
is conceptualized in such a way that it can be used as a selfcontained textbook or as an integral part of the online
learning course.
A number of individuals deserve to be singled out as
their support was crucial in this development and is highly
appreciated: Dr. Roshan Shrestha, Dr. Doulaye Koné, Dr.
Frank Rijsberman and Dr. Brian Arbogast (BMGF), and
Dr. Wim Duven and Jetze Heun (UNESCO-IHE). The
book was edited by Peter Stroo, Hans Emeis, Claire Taylor,
Michelle Jones, and Maggie Smith. The credit for the
content goes to all the authors, reviewers and enthusiastic
group of editors. Further, I acknowledge the contributors
who allowed their data, images and photographs to be used
in this book and the course.
Finally, I hope that this book and the training
materials will be useful in your research or practical work,
be it at a laboratory-, pilot- or full-scale wastewater
treatment plant.
Prof. Dr. Damir Brdjanovic
Professor of Sanitary Engineering
Table of contents
1. INTRODUCTION
1
Mark C.M. van Loosdrecht, Per H. Nielsen, Carlos M. Lopez-Vazquez
and Damir Brdjanovic (aut.)
2. ACTIVATED SLUDGE ACTIVITY TESTS
7
Carlos M. Lopez-Vazquez, Laurens Welles, Tommaso Lotti, Elena Ficara,
Eldon R. Rene, Tessa P.H. van den Brand, Damir Brdjanovic
and Mark C.M. van Loosdrecht (aut.)
Yves Comeau, Piet N.L. Lens and Nancy G. Love (rev.)
2.1 INTRODUCTION
2.2 ENHANCED BIOLOGICAL PHOSPHORUS REMOVAL
2.2.1 Process description
2.2.2 Experimental set-up
2.2.2.1 Reactors
2.2.2.2 Activated sludge sample collection
2.2.2.3 Activated sludge sample preparation
2.2.2.4 Substrate
2.2.2.5 Analytical procedures
2.2.2.6 Parameters of interest
2.2.3 EBPR batch activity tests: Preparation
2.2.3.1 Apparatus
2.2.3.2 Materials
2.2.3.3 Media preparation
2.2.3.4 Material preparation
2.2.3.5 Activated sludge preparation
2.2.4 Batch activity tests: Execution
2.2.4.1 Anaerobic EBPR batch activity tests
2.2.4.2 Anoxic EBPR batch tests
2.2.4.3 Aerobic EBPR batch tests
2.2.5 Data analysis
2.2.5.1 Estimation of stoichiometric parameters
2.2.5.2 Estimation of kinetic parameters
2.2.6 Data discussion and interpretation
2.2.6.1 Anaerobic batch activity tests
2.2.6.2 Aerobic batch activity tests
2.2.6.3 Anoxic batch activity tests
2.2.7 Example
2.2.7.1 Description
2.2.7.2 Data analysis
2.2.8 Additional considerations
2.2.8.1 GAO occurrence in EBPR systems
2.2.8.2 The effect of carbon source
2.2.8.3 The effect of temperature
2.2.8.4 The effect of pH
2.2.8.5 Denitrification by EBPR cultures
2.2.8.6 Excess and shortage of intracellular compounds
2.2.8.7 Excessive aeration
2.2.8.8 Shortage of essential ions
2.2.8.9 Toxicity/inhibition
2.3 BIOLOGICAL SULPHATE REDUCTION
2.3.1 Process description
2.3.2 Sulphide speciation
2.3.3 Effects of environmental and operating conditions on SRB
2.3.3.1 Carbon source
2.3.3.2 COD to SO42- ratio
2.3.3.3 Temperature
2.3.3.4 pH
7
9
9
11
11
16
16
17
19
22
22
22
24
24
25
28
29
30
33
34
36
36
41
42
42
45
46
47
47
47
51
51
51
51
52
52
52
53
53
53
54
54
56
57
57
58
58
59
2.3.3.5 Oxygen
2.3.4 Experimental set-up
2.3.4.1 Estimation of volumetric and specific rates
2.3.4.2 The reactor
2.3.4.3 Mixing
2.3.4.4 pH control
2.3.4.5 Temperature control
2.3.4.6 Sampling and dosing ports
2.3.4.7 Sample collection
2.3.4.8 Media
2.3.5 Analytical procedures
2.3.5.1 CODorganics and CODtotal
2.3.5.2 Sulphate
2.3.5.3 Sulphide
2.3.6 SRB batch activity tests: preparation
2.3.6.1 Apparatus
2.3.6.2 Materials
2.3.6.3 Media
2.3.6.4 Material preparation
2.3.6.5 Mixed liquor preparation
2.3.6.6 Sample collection and treatment
2.3.7 Batch activity tests: execution
2.3.8 Data analysis
2.3.8.1 Mass balances and calculations
2.3.8.2 Data discussion and interpretation
2.3.9 Example
2.3.10 Practical recommendations
2.4 BIOLOGICAL NITROGEN REMOVAL
2.4.1 Process description
2.4.1.1 Nitrification
2.4.1.2 Denitrification
2.4.1.3 Anaerobic ammonium oxidation (Anammox)
2.4.2 Process-tracking alternatives
2.4.2.1 Chemical tracking
2.4.2.2 Titrimetric tracking
2.4.2.3 Manometric tracking
2.4.3 Experimental set-up
2.4.3.1 Reactors
2.4.3.2 Instrumentation for titrimetric tests
2.4.3.3 Instrumentation for manometric tests
2.4.3.4. Activated sludge sample collection
2.4.3.5 Activated sludge sample preparation
2.4.3.6 Substrate
2.4.3.7 Analytical procedures
2.4.3.8 Parameters of interest
2.4.3.9 Type of batch tests
2.4.4 Nitrification batch activity tests: Preparation
2.4.4.1 Apparatus
2.4.4.2 Materials
2.4.4.3 Media preparation
2.4.5 Nitrification batch activity tests: Execution
2.4.6 Denitrification batch activity tests: Preparation
2.4.6.1 Apparatus
2.4.6.2 Materials
2.4.6.3 Working solutions
2.4.6.4 Materials preparation
2.4.7 Denitrification batch activity tests: Execution
2.4.8 Anammox batch activity tests: Preparation
59
60
60
60
61
61
61
62
62
62
63
63
64
64
65
65
65
65
66
67
68
68
69
69
70
70
72
73
73
74
75
76
76
77
77
78
79
79
79
80
81
82
82
83
83
86
86
86
86
86
87
92
92
93
93
93
93
99
2.4.8.1 Apparatus
2.4.8.2 Materials
2.4.8.3 Working solutions
2.4.8.4 Materials preparation
2.4.9 Anammox batch activity tests: Execution
2.4.10 Examples
2.4.10.1 Nitrification batch activity test
2.4.10.2 Denitrification batch activity test
2.4.10.3 Anammox batch activity test
2.4.11 Additional considerations
2.4.11.1 Presence of other organisms
2.4.11.2 Shortage of essential micro- and macro-nutrients
2.4.11.3 Toxicity or inhibition effects
2.4.11.4 Effects of carbon source on denitrification
2.5 AEROBIC ORGANIC MATTER REMOVAL
2.5.1 Process description
2.5.2 Experimental set-up
2.5.2.1 Reactors
2.5.2.2 Activated sludge sample collection
2.5.2.3 Activated sludge sample preparation
2.5.2.4 Media
2.5.2.5 Analytical tests
2.5.2.6 Parameters of interest
2.5.3 Aerobic organic matter batch activity tests: Preparation
2.5.3.1 Apparatus
2.5.3.2 Materials
2.5.3.3 Working solutions
2.5.3.4 Material preparation
2.5.3.5 Activated sludge preparation
2.5.4 Aerobic organic matter batch activity tests: Execution
2.5.5 Data analysis
2.5.6 Example
2.5.6.1 Description
2.5.6.2 Data analysis
2.5.7 Additional considerations and recommendations
2.5.7.1 Simultaneous storage and microbial growth
2.5.7.2 Lack of nutrients
2.5.7.3 Toxicity or inhibition
99
99
99
100
100
103
103
105
107
109
109
109
110
110
111
111
112
112
112
113
113
114
114
115
115
115
115
116
117
117
118
119
119
119
121
121
121
121
3. RESPIROMETRY
133
Henry Spanjers and Peter A. Vanrolleghem (aut.)
George A. Ekama and M. Spérandio (rev.)
3.1 INTRODUCTION
3.1.1 Basics of respiration
3.1.2 Basics of respirometry
3.2 GENERAL METHODOLOGY OF RESPIROMETRY
3.2.1 Basics of respirometric methodology
3.2.2 Generalized principles: beyond oxygen
3.2.2.1 Principles based on measuring in the liquid phase
3.2.2.2 Principles based on measuring during the gas phase
3.3 EQUIPMENT
3.3.1 Equipment for anaerobic respirometry
3.3.1.1 Biogas composition
3.3.1.2 Measuring the gas flow
3.3.2 Equipment for aerobic and anoxic respirometry
3.3.2.1 Reactor
3.3.2.2 Measuring arrangement
3.3.2.3 Practical implementation
3.4 WASTEWATER CHARACTERIZATION
3.4.1 Biomethane potential (BMP)
3.4.1.1 Purpose
3.4.1.2 General
133
134
135
136
136
136
136
138
141
141
141
142
143
143
143
144
150
150
150
150
3.4.1.4 Data processing
3.4.1.5 Recommendations
3.4.2 Biochemical oxygen demand (BOD)
3.4.2.1 Purpose
3.4.2.2 General
3.4.2.3 Test execution
3.4.3 Short-term biochemical oxygen demand (BODst)
3.4.3.1 Test execution
3.4.3.2 Calculations
3.4.4 Toxicity and inhibition
3.4.4.1 Purpose
3.4.4.2 Test execution
3.4.4.3 Calculations
3.4.4.4 Biodegradable toxicants
3.4.5 Wastewater fractionation
3.4.5.1 Readily biodegradable substrate (SB)
3.4.5.2 Slowly biodegradable substrate (XCB)
3.4.5.3 Heterotrophic biomass (XOHO)
3.4.5.4 Autotrophic (nitrifying) biomass (XANO)
3.4.5.5 Ammonium (SNHx)
3.4.5.6 Organic nitrogen fractions (XCB,N and SB,N)
3.5 BIOMASS CHARACTERIZATION
3.5.1 Volatile suspended solids
3.5.2 Specific methanogenic activity (SMA)
3.5.2.1 Purpose
3.5.2.2 General
3.5.2.3 Test execution
3.5.2.4 Data processing
3.5.3 Specific aerobic and anoxic biomass activity
3.5.3.1 Maximum specific nitrification rate (AUR)
3.5.3.2 Maximum specific aerobic heterotrophic
respiration rate (OUR)
3.5.3.3 Maximum specific denitrification rate (NUR)
151
151
152
152
152
153
157
158
160
160
160
160
161
162
163
166
167
168
168
168
168
169
169
169
169
169
169
170
171
171
4. OFF-GAS EMISSION TESTS
177
173
173
Kartik Chandran, Eveline I.P. Volcke, Mark C.M. van Loosdrecht (aut.)
Peter A. Vanrollegem and Sylvie Guillot (rev.)
4.1 INTRODUCTION
4.2 SELECTING THE SAMPLING STRATEGY
4.2.1 Plant performance
4.2.2 Seasonal variations in emissions
4.2.3 Sampling objective
4.3 PLANT ASSESSMENT AND DATA COLLECTION
4.3.1 Preparation of a sampling campaign
4.3.2 Sample identification and data sheet
4.3.3 Factors that can limit the validity of the results
4.3.4 Practical advice for analytical measurements
4.3.5 General methodology for sampling
4.3.6 Sampling in the framework of the off-gas measurements
4.3.7 Testing and measurements protocol
4.4 EMISSION MEASUREMENTS
4.5 N2O MEASUREMENT IN OPEN TANKS
4.5.1 Protocol for measuring the surface flux of N2O
4.5.1.1 Equipment, materials and supplies
4.5.1.2 Experimental procedure
4.5.1.3 Sampling methods for nitrogen GHG emissions
4.5.1.4 Direct measurement of the liquid-phase N2O content
4.6 MEASUREMENT OF OFF-GAS FLOW IN OPEN TANKS
4.6.1 Protocol for aerated or aerobic zone
4.6.2 Protocol for non-aerated zones
4.7 AQUEOUS N2O and CH4 CONCENTRATION DETERMINATION
177
178
178
178
179
179
179
180
181
181
182
183
185
185
186
188
188
188
189
191
191
192
192
192
4.7.1 Measurement protocol for dissolved N2O measurement
using polarographic electrodes
4.7.1.1 Equipment
4.7.1.2 Experimental procedure
4.7.2 Measurement protocol for dissolved gasses using
gas chromatography
4.7.3 Measurement protocol for dissolved gas measurement
by the salting-out method
4.7.3.1 Equipment
4.7.3.2 Sampling procedure
4.7.3.3 Measurement procedure
4.7.3.4 Calculations
4.7.4 Measurement protocol for dissolved gas measurement
by the stripping method
4.7.4.1 Operational principle
4.7.4.2 Equipment
4.7.4.3 Calibration batch test
4.7.4.4 Measurement accuracy
4.7.4.5 Calculation of the N2O formation rate in the stripping device
4.8 DATA ANALYSIS AND PROCESSING
4.8.1 Determination of fluxes
4.8.2 Determination of aggregated emission fractions
4.8.3 Calculation of the emission factors
5. DATA HANDLING AND PARAMETER ESTIMATION
193
193
193
194
194
195
195
195
196
196
196
197
198
198
198
199
199
199
200
201
Gürkan Sin and Krist V. Gernaey (aut.)
Sebastiaan C.F. Meijer and Juan A. Baeza (rev.)
5.1 INTRODUCTION
5.2 THEORY AND METHODS
5.2.1 Data handling and validation
5.2.1.1 Systematic data analysis for biological processes
5.2.1.2 Degree of reduction analysis
5.2.1.3 Consistency check of experimental data
5.2.2 Parameter estimation
5.2.2.1 Manual trial and error method
5.2.2.2 Formal statistics methods
5.2.3 Uncertainty analysis
5.2.3.1 Linear error propagation
5.2.3.2 The Monte Carlo method
5.2.4 Local sensitivity analysis and identifiability analysis
5.2.4.1 Local sensitivity analysis
5.2.4.2 Identifiability analysis using the collinearity index
5.3 METHODOLOGY AND WORKFLOW
5.3.1 Data consistency check using elemental balance and
a degree of reduction analysis
5.3.2 Parameter estimation workflow for non-linear least
squares method
5.3.3 Parameter estimation workflow for the bootstrap method
5.3.4 Local sensitivity and identifiability analysis workflow
5.3.5 Uncertainty analysis using the Monte Carlo method and
linear error propagation
5.4 ADDITIONAL EXAMPLES
5.5 ADDITIONAL CONSIDERATIONS
201
202
202
202
203
204
205
205
205
209
209
209
210
210
210
211
6. SETTLING TESTS
235
211
212
212
213
213
214
232
Elena Torfs, Ingmar Nopens, Mari K.H. Winkler, Peter A. Vanrolleghem,
Sophie Balemans and Ilse Y. Smets (aut.)
Glenn T. Daigger and Imre Takács (rev.)
6.1 INTRODUCTION
6.2 MEASURING SLUDGE SETTLEABILITY IN SSTs
6.2.1 Sludge settleability parameters
6.2.1.1 Goal and application
6.2.1.2 Equipment
6.2.1.3 The sludge volume index (SVI)
6.2.1.4 The diluted sludge volume index (DSVI)
6.2.1.5 The stirred specific volume index (SSVI3.5)
6.2.2 The batch settling curve and hindered settling velocity
6.2.2.1 Goal and application
6.2.2.2 Equipment
6.2.2.3 Experimental procedure
6.2.2.4 Interpreting a batch settling curve
6.2.2.5 Measuring the hindered settling velocity
6.2.3 vhs-X relation
6.2.3.1 Goal and application
6.2.3.2 Equipment
6.2.3.3 Experimental procedure
6.2.3.4 Determination of the zone settling parameters
6.2.3.5 Calibration by empirical relations based on SSPs
6.2.4 Recommendations for performing batch settling tests
6.2.4.1 Shape and size of the batch reservoir
6.2.4.2 Sample handling and transport
6.2.4.3 Concentration range
6.2.4.4 Measurement frequency
6.2.5 Recent advances in batch settling tests
6.3 MEASURING FLOCCULATION STATE OF ACTIVATED SLUDGE
6.3.1 DSS/FSS test
6.3.1.1 Goal and application
6.3.1.2 Equipment
6.3.1.3 DSS test
6.3.1.4 FSS test
6.3.1.5 Interpretation of a DSS/FSS test
6.3.2 Recommendations
6.3.2.1 Flocculation conditions
6.3.2.2 Temperature influence
6.3.2.3 Supernatant sampling
6.3.3 Advances in the measurement of the flocculation state
6.4 MEASURING THE SETTLING BEHAVIOUR OF GRANULAR SLUDGE
6.4.1 Goal and application
6.4.2 Equipment
6.4.3 Density measurements
6.4.4 Granular biomass size determination
6.4.4.1 Sieving
6.4.4.2 Image analyser
6.4.5 Calculating the settling velocity of granules
6.4.6 Recommendations
6.4.6.1 Validation of results
6.4.6.2 Application for flocculent sludge
6.5 MEASURING SETTLING VELOCITY DISTRIBUTION IN PSTs
6.5.1 Introduction
6.5.2 General principle
6.5.3 Sampling and sample preservation
6.5.4 Equipment
6.5.5 Analytical protocol
6.5.6 Calculations and result presentation
6.5.6.1 Mass balance check
6.5.6.2 Calculation of the settling velocity distribution
6.5.6.3 Recommendations
237
237
237
238
238
238
239
239
240
241
241
241
242
242
243
244
245
245
245
245
245
245
246
246
246
246
246
247
248
249
249
249
249
250
250
250
251
251
252
252
253
253
254
254
255
255
255
255
256
256
257
258
258
258
259
7. MICROSCOPY
263
Jeppe L. Nielsen, Robert J. Seviour and Per H. Nielsen (aut.)
Jiři Wanner (rev.)
235
236
237
237
7.1 INTRODUCTION
7.2 THE LIGHT MICROSCOPE
7.2.1 Standard applications of light microscopy
263
263
265
7.2.2 Low power objective
7.2.3 High power objective
7.2.4 Immersion objective
7.2.5 Important considerations
7.2.6 Bright-field and dark-field illumination
7.2.7 Fluorescence microscopy
7.2.8. Confocal laser scanning microscopy
7.3 MORPHOLOGICAL INVESTIGATIONS
7.3.1 Microscopic ‘identification’ of filamentous microorganisms
7.3.2 ‘Identification’ of protozoa and metazoa
7.4 EXAMINING ACTIVATED SLUDGE SAMPLES MICROSCOPICALLY
7.4.1 Mounting the activated sludge sample
7.4.2 Gram staining
7.4.2.1 Reagents and solutions for Gram staining
7.4.2.2 Procedure
7.4.3 Neisser staining
7.4.3.1 Reagents and solutions for Neisser staining
7.4.3.2 Procedure
7.4.4 DAPI staining
7.4.4.1 Reagents and solutions for DAPI staining
7.4.4.2 Procedure
7.4.5 CTC staining
7.4.5.1 Reagents and solutions for CTC staining
7.4.5.2 Procedure
7.5 FLUORESCENCE in situ HYBRIDIZATION
7.5.1 Reagents and solutions for FISH
7.5.2 Procedure
7.6 COMBINED STAINING TECHNIQUES
7.6.1 FISH-DAPI staining
7.6.1.1 Reagents and solutions for DAPI staining
7.6.1.2 Procedure
7.6.2 FISH-PHA staining
7.6.2.1 Reagents and solutions for PHA staining
7.6.2.2 Procedure
265
265
265
266
266
267
269
269
270
271
272
272
273
273
274
274
274
275
275
275
275
276
276
276
276
277
278
280
281
281
281
282
282
282
8. MOLECULAR METHODS
285
Søren M. Karst, Mads Albertsen, Rasmus H. Kirkegaard, Morten S. Dueholm
and Per H. Nielsen (aut.)
Holger Daims (rev.)
8.1 INTRODUCTION
8.2 EXTRACTION OF DNA
8.2.1 General considerations
8.2.2 Sampling
8.2.3 DNA extraction
8.2.3.1 Cell lysis
8.2.3.2 Nuclease activity inhibition and protein removal
8.2.3.3 Purification
8.2.3.4 Elution and storage
8.2.4 Quantification and integrity
8.2.5 Optimised DNA extraction from wastewater activated sludge
8.2.5.1 Materials
8.2.5.2 DNA extraction
8.3 REAL-TIME QUANTITATIVE PCR (qPCR)
8.3.1 General considerations
8.3.2 Materials
8.3.3 Methods
8.3.4 Data handling
8.3.5 Data output and interpretation
8.3.6 Troubleshooting
8.3.7 Example
8.3.7.1 Samples
8.3.7.2 qPCR reaction setup
285
286
286
286
286
286
287
287
287
287
288
288
288
289
289
291
292
294
294
295
295
295
296
8.3.7.3 Results
8.4 AMPLICON SEQUENCING
8.4.1 General considerations
8.4.2 The 16S rRNA gene as a phylogenetic marker gene
8.4.3 PCR amplification
8.4.3.1 PCR reaction
8.4.3.2 PCR biases
8.4.3.3 Primer choice
8.4.4 DNA sequencing
8.4.4.1 Sequencing platform
8.4.4.2 Sequencing depth
8.4.5 Bioinformatic processing
8.4.5.1 Available software
8.4.5.2 Raw data
8.4.5.3 Quality scores and filtering
8.4.5.4 Merging paired end-reads
8.4.5.5 OTU clustering
8.4.5.6 Chimera detection and removal
8.4.5.7 Taxonomic classification
8.4.5.8 The OTU table
8.4.6 Data analysis
8.4.6.1 Defining the goal of the data analysis
8.4.6.2 Data validation and sanity check
8.4.6.3 Communities or individual species?
8.4.6.4 Identifying core and transient species
8.4.6.5 Explorative analysis using multivariate statistics
8.4.6.6 Correlation analysis
8.4.6.7 Effect of treatments on individual species
8.4.7 General observations
8.4.7.1 A relative analysis
8.4.7.2 Copy number bias
8.4.7.3 Primer bias
8.4.7.4 Standardization
8.4.7.5 Impact of the method
8.4.8 Protocol: Illumina V1-3 16S rRNA amplicon libraries
8.4.8.1 Apparatus
8.4.8.2 Materials
8.4.8.3 Protocol
8.4.9 Interpretation and troubleshooting
8.4.9.1 Sample DNA quality control and dilution
8.4.9.2 Library PCR
8.4.9.3 Library cleanup
8.4.9.4 Library quality control
8.4.9.5 Library pooling
8.4.9.6 Pool quality control and dilution
8.4.9.7 Storage
8.4.10 Protocol: Illumina V1-3 16S amplicon sequencing
8.4.10.1 Apparatus
8.4.10.2 Reagents
8.4.10.3 Protocol
8.4.10.4 Interpretation and troubleshooting
8.4.11 Design of Illumina 16S amplicon sequencing adaptors
8.5 OTHER METHODS
LIST OF SYMBOLS AND ABBREVIATIONS
296
297
297
297
299
299
300
300
301
301
301
301
301
302
303
303
303
304
304
304
304
304
305
305
306
306
307
307
307
307
307
307
308
308
308
308
308
309
311
311
312
312
313
314
314
314
314
314
314
314
315
317
319
1
INTRODUCTION
Authors:
Mark C.M. van Loosdrecht
Per H. Nielsen
Carlos M. Lopez-Vazquez
Damir Brdjanovic
Wastewater treatment forms a crucial link in the services
that the sanitation sector delivers to society. For
centuries, sanitation largely consisted of transporting
fresh, clean water to the cities, and using this water to
transport the waste out of the city and discharge it into
the natural environment. However, with the increase in
human populations in cities as a result of the industrial
revolution in the 19th century, this could no longer be
maintained. The occurrence of epidemic diseases
facilitated the development of wastewater treatment
facilities and their implementation since the early 20th
century. This development has been largely an empirical
activity with theoretical approaches following
experimental observations (Figure 1.1).
Figure 1.1 Noyes Laboratory on the campus of the University of Illinois in
Urbana was arguably the most important in promoting research in
wastewater in the early 20th century (photo: University of Illinois, 1902).
The discovery and development of activated sludge
technology (described in detail in Jenkins and Wanner,
2014) was crucial as it triggered the rapid development
and application of various analytical and experimental
methods. Experimental work in the Lawrence
Experimental Station in Massachusetts, USA, which at
that time (1912) was a unique facility aimed at the
experimental verification of different possible
wastewater treatment procedures, inspired Gilbert
Fowler to request Edward Ardern and William Lockett to
repeat the experiments with wastewater aeration in the
UK that he had seen in the USA. In 1913 and 1914
Lockett and Ardern carried out lab-scale experiments at
the Manchester - Davyhulme wastewater treatment plant
(Figure 1.2). Glass bottles were used to represent labscale aeration basins ‘fed’ by sewage from different
districts of Manchester. Contrary to the experiments that
Fowler saw in Massachusetts, in the Manchester aeration
tests the sediment that remained after decantation was left
in the bottle and a new dose of sewage was added to the
sediment for the next batch. Lockett and Ardern soon
found that the amount of the sediment increased with the
increasing number of batches. At the same time the
aeration time necessary for ‘full oxidation’ of sewage
(full oxidation was a term used to describe the removal
of degradable organics and for complete nitrification)
was reduced. By using this technique of repeated batch
aeration with the sediment remaining in the bottle,
Lockett and Ardern were able to shorten the required
aeration time for ‘full oxidation’ from a few weeks to less
than one day, which made the process technically
© 2016 Mark C.M. van Loosdrecht et al. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen. C.M. Lopez-Vazquez and D. Brdjanovic. ISBN:
9781780404745 (Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
2
feasible. The sediment formed during the aeration of
sewage was called activated sludge due to its appearance
and activity. Lockett and Ardern published their results
in a famous series of three papers (Ardern and Lockett
1914a, 1914b, 1915). This was the ‘birth’ of activated
sludge, which is today the workhorse of wastewater
treatment and the most widely applied sewage treatment
technology in the world.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Examples are the commonly used chemical or biological
oxygen demand tests. The iconic ‘Standard Methods for
the Examination of Water and Wastewater’ (APHA et al.,
2012, Figure 1.3) has for generations of sanitary
engineers been the resource for analysing their
experimental systems and full-scale operations. These
methods focus heavily on the chemical characterization
and measurement of specific microorganisms.
Figure 1.2 The Davyhulme Sewage Works Laboratory, where the activated
sludge process was developed in the early 20th century (photo: United
Utilities).
Wastewater engineering is a profession that is
extremely experiment-based, and therefore it has always
had the need to develop and standardise methods. This
seemingly simple activity is strongly hampered by two
factors, namely: (i) wastewater engineering is a typical
interdisciplinary activity where chemical engineers, civil
engineers, microbiologists and chemists interact to
develop and understand the processes; the challenge here
is to integrate methods and approaches from these
disciplines, and, (ii) in addition, wastewater and its
treatment processes are by their nature difficult to define
with exactitude. It is for instance virtually impossible to
measure all the individual compounds in the wastewater
itself. Identifying all the relevant microorganisms in the
processes has long been impossible and is still a
complicated challenge. Defining all the potentially
occurring chemical conversions is, due to the myriad of
chemicals present, again an almost impossible task.
Due to the undefined nature of the experimental
system, research has tended to progress slowly and it
heavily depends on standardised methods that may not be
exact but, when used in a standardised way, are very
helpful and useful to compare experimental results.
Figure 1.3 The Standard Methods for the Examination of Water and
Wastewater. The first edition appeared in 1905 (image: APHA et al., 2012).
Societal demands on the efficiency of wastewater
treatment plants have advanced, moving from public
health protection to water resources and environmental
protection and nowadays to integrated resource and
energy recovery. Therefore the need to accurately
characterize the microbial processes in the wastewater
treatment processes has increased over recent decades.
Certainly, it is a challenge to develop standardized
methods for experimental work that can be easily
repeated in different laboratories. In many cases, the
exact handling is important, but it is not easy to be written
down in a practical protocol.
INTRODUCTION
Therefore, to avoid these problems, it was decided to
develop not only a book describing all the experimental
methods but also a video catalogue with the methods
described in this book actually being demonstrated in the
laboratory. This book and its associated video-based
material are designed to support the research and
development field with a manual for characterizing the
biological processes in wastewater treatment. The editors
have decided in this first edition of the ‘Experimental
methods in wastewater treatment’ book to focus on the
activated sludge process since this is worldwide by far
the most applied technology. Nevertheless, most of the
methods presented in this book can also be applicable to
biofilm-based technologies or anaerobic digestion
processes.
The decision to focus on experimental methods
related to the activated sludge process has resulted in
seven chapters describing the key experimental methods.
The content and focus of these chapters are summarised
in Table 1.1. Activated sludge consists of a myriad of
microorganisms, converting a range of important
compounds (organic matter, oxygen, nitrogen and
phosphate compounds). The first three chapters focus on
characterizing the conversion capacities of the microbial
communities for the major microbial processes. A
distinction has been made between full liquid-phasebased methods and methods where the conversion are
characterized by measuring the respiration of the
organisms, usually gas-phase measurements. Since there
is an increasing focus on and interest in assessing the
environmental impact of wastewater treatment plants, a
separate chapter has been added for measuring
greenhouse gas emissions from wastewater treatment
plants. These chapters are followed by a chapter
describing data handling techniques. Measurements
often, certainly from full-scale or pilot plants, have
relatively large uncertainties. With adequate data
handling techniques the measurements can be used to
derive associated (difficult to directly measure) process
data or to minimise their uncertainty.
Activated sludge processes mainly depend on settling
of the flocculent sludge to separate the biomass from the
cleaned wastewater. This is often the Achilles heel of the
treatment process and a key factor in the process design.
One chapter is therefore devoted to characterization of
the sludge settling properties.
As said earlier, microorganisms are the workhorses in
the activated sludge process. Therefore the microscope is
unavoidably the main technique to observe them directly,
not only for individual organisms but also for the floc
3
morphology related to settling characteristics. For a long
time the microscope has been the main method of choice
when observing which bacteria are present in activated
sludge. However, although very helpful, it cannot show
the full complexity of the microbial community. The last
decade’s advance in molecular DNA-based techniques
has revolutionized the way one can observe
microorganisms. These generic novel methods are
described in the final chapter of this book.
Within the chapters the authors have tried to describe
especially those methods that are experimentally
complex and not standard analytical procedures.
Therefore, standard analytical methods for e.g. organic
matter, ammonium, phosphate etc. are not described in
detail. On the other hand, it was also decided to include
some analytical techniques recently developed and/or
improved that are becoming frequently used but are
scattered across scientific literature (e.g. glycogen and
poly-hydroxy-alkanoates determination). In addition,
methods that could be of academic interest but currently
have limited practical application have not been included
in detail in the text.
In terms of symbols and notation, an attempt has been
made to standardize them as much as possible. While this
was achieved at the chapter level, full standardization
was not possible across all the chapters due to their
diverse nature and heterogeneity of items as well as lack
of global agreement on the use of symbols and notations,
although the most common guidelines were quite closely
followed (e.g. Corominas et al., 2010).
The book is conceptualized so as to satisfy users with
high demands who are able to handle complex analytical
and experimental equipment. However, the content is
equally suited to the requirements of less advanced
laboratories and less experienced experimenters; in
particular, the complementary, freely available video
materials address the execution of experiments in more
challenging environments, such as those usually
prevailing in most less developed countries.
"To measure is to know."
Lord Kelvin
4
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 1.1 A simplified overview of the experimental methods presented in the book per process of interest.
Process
Introduction
Organic matter
removal
Activated sludge
activity tests
Kinetics
AOO and NOO
activity
Kinetics
Stoichiometry
Nitrification
Overview and
rationale to
experimental
methods
Denitrification
Anammox
EBPR
Anaerobic
treatment
Respirometry
Biochemical
oxygen demand
(BOD)
Short-term
biochemical
oxygen demand
Wastewater
characterization
and fractionation
Biomass
characterization
Toxicity and
inhibition
Wastewater
characterization
and fractionation
Biomass
characterization
AOO and NOO
activity
Toxicity and
inhibition
Kinetics
Stoichiometry
Denitrification
over NO2 and NO3
Denitrification
on RBCOD and
SBCOD
Stoichiometry
Kinetics
AMX activity
Kinetics
Stoichiometry
Denitrification
over NO2 and NO3
Toxicity and
inhibition
Stoichiometry
Kinetics
PAO, GAO, and
Aerobic kinetics
DPAO activity
Kinetics
Stoichiometry
SRB activity
Kinetics
Stoichiometry
Chapter
Off-gas emission Data handling and Settling tests
tests
parameter
estimation
Sampling
methods for
nitrogen GHG
emissions
Methods for
off-gas
measurements
Aqueous N2O and
CH4 concentration
determination
methods
Gas
measurement
methods in open
tanks
Microscopy
Light
microscopy
distributions in
Confocal
primary settling microscopy
tanks
Morphological
investigations
Sludge
settleability in
Staining
secondary settling techniques
tanks
Fluorescence in
situ
Flocculation
Hybridization
properties
(FISH)
Settling
behaviour of
Combined
granular sludge staining
techniques
Molecular
methods
Settling velocity
Data handling
and validation
Parameter
estimation
Uncertainty
analysis
Local sensitivity
analysis and
identifiability
analysis
DNA extraction
Real-time
quantitative PCR
Amplicon
sequencing
and
stoichiometry
Toxicity and
inhibition
Specific
methanogenic
activity
Biomethane
potential
Toxicity and
inhibition
Kinetics
Stoichiometry
Settling
AMX Anammox organisms
AOO Ammonium oxidizing organisms
CH4 Methane
DNA Deoxyribonucleic acid
DPAO Denitrifying poly-phosphate accumulating organisms
EBPR Enhanced biological phosphorus removal
FISH Fluorescence in situ hybridization
GAO Glycogen accumulating organisms
GHG Greenhouse gas emissions
N2O Nitrous oxide
NO2 Nitrite
NO3 Nitrate
NOO Nitrite oxidizing organisms
PAO Poly-phosphate accumulating organisms
PCR Polymerase chain reaction
RBCOD Readily biodegradable COD also known as readily biodegradable organics
SBCOD Slowly biodegradable COD also known as slowly biodegradable organics
SRB Sulphate reducing bacteria or SRO Sulphate reducing organism
INTRODUCTION
5
Figure 1.4 The mission of UNESCO-IHE is to contribute to the education and training of professionals, to expand the knowledge base through research and
to build the capacity of sector organizations, knowledge centres and other institutions active in the fields of water, the environment and infrastructure in
developing countries and countries in transition. The photos depict the illustrative example of the Institute's latest project in Cuba where the laboratory of the
Instituto de Investigaciones para la Industria Alimenticia (IIIA) in Havana has been equipped with new state-of-the-art technology and where the local staff
has been trained on how to operate the equipment and prepare and carry out experimental work (photo: Brdjanovic, 2015).
References
American Public Health Association (APHA), American Water
Works Association (AWWA), and Water Environment
Federation (WEF) (2012). Standard Methods for the
Examination of Water and Wastewater, 22nd Edition. New
York. ISBN 9780875530130.
Ardern, E., Lockett, W.T. (1914a) Experiments on the Oxidation of
Sewage without the Aid of Filters. J. Soc. Chem. Ind., 33: 523.
Ardern, E., Lockett, W.T. (1914b) Experiments on the Oxidation of
Sewage without the Aid of Filters, Part II. J. Soc. Chem. Ind.,
33: 1122.
Ardern, E., Lockett, W.T. (1915) Experiments on the Oxidation of
Sewage without the Aid of Filters, Part III. J. Soc. Chem. Ind.,
34: 937.
Corominas, L.L., Rieger, L., Takács, I., Ekama, A.G., Hauduc, H.,
Vanrolleghem, P.A., Oehmen, A., Gernaey, K.V., van
Loosdrecht, M.C.M., Comeau Y. (2010). New framework for
standardized notation in wastewater treatment modelling.
Water Sci Technol. 61(4): 841-57.
Jenkins, D. and Wanner, J. Eds. (2014) 100 years of activated sludge
and
counting.
IWA
Publishing,
London,
ISBN
9781780404936, pg. 464.
The section on activated sludge historical development
presented in this chapter is adapted from Jenkins and Wanner
(2014).
6
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2
ACTIVATED SLUDGE ACTIVITY TESTS
Authors:
Reviewers:
Carlos M. Lopez-Vazquez
Laurens Welles
Tommaso Lotti
Elena Ficara
Eldon R. Rene
Tessa P.H. van den Brand
Damir Brdjanovic
Mark C.M. van Loosdrecht
Yves Comeau
Piet N.L. Lens
Nancy G. Love
2.1 INTRODUCTION
Different conditions and factors affect the degree and rate
(speed) at which the compounds and contaminants of
concern are removed by microbial populations in
biological wastewater treatment systems. Certainly, the
plant configuration and operational conditions play a
major role in the prevalence of specific microbial
populations and their activities, but factors as diverse and
broad as wastewater characteristics and environmental
and climate conditions have a strong influence as well.
Eventually, in any biological wastewater treatment
system, there will be a need to assess, define and
understand the plant performance with regard to the
removal of certain contaminants and the response of the
sludge to inhibitory or toxic compounds of interest.
Moreover, from a modelling perspective it is also of
interest to assess and determine the stoichiometry and
kinetic rates of the conversion processes performed by
specific microbial populations (e.g. ordinary
heterotrophic organisms: OHOs; denitrifying ordinary
heterotrophic organisms: dOHOs; ammonium-oxidizing
organisms: AOOs; nitrite-oxidizing organisms: NOOs;
phosphate-accumulating organisms: PAOs; sulphatereducing bacteria: SRB, also identified as sulphatereducing organisms, SRO (Corominas et al., 2010); or,
anaerobic ammonium-oxidizing organisms: anammox.
Thereby, the execution of batch activity tests can be
rather useful to: (i) study the biodegradability of a given
wastewater stream (municipal or industrial), (ii)
determine the stoichiometric and kinetic parameters
involved in the conversion of a specific compound, (iii)
study the potential interactions (e.g. symbiosis and
competition) between microbial populations and (iv)
assess the potential inhibitory or toxic effects of certain
wastewaters, compounds or substances.
The nature and type of the batch activity tests can
differ depending upon the compounds of interest and the
metabolism and physiology of the microbial populations
involved in the removal or conversion processes. For
instance, they can range from relatively simple aerobic
tests where organic matter removal by OHOs is measured
to more complex alternating anaerobic-anoxic-aerobic
© 2016 Carlos M. Lopez-Vazquez et al. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen. C.M. Lopez-Vazquez and D. Brdjanovic. ISBN:
9781780404745 (Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
8
batch tests
presence of
nitrite and
performing
(EBPR).
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
to assess the activity of PAOs under the
different electron acceptors (such as nitrate,
oxygen) from activated sludge systems
enhanced biological phosphorus removal
This chapter presents an overview of the most
common batch activity tests and protocols and their
execution with the aim of assessing the conversion
processes involved in: (i) enhanced biological
phosphorus removal by PAOs under alternating
anaerobic-aerobic conditions, (ii) denitrification via
nitrate or nitrite by PAOs, (iii) reduction of sulphate by
SRBs, (iv) removal of organics under aerobic conditions
by OHOs, (v) denitrification by dOHOs using nitrate or
nitrite as final electron acceptor, (vi) oxidation of
ammonia and nitrite by AOOs and NOOs under aerobic
conditions and (vii) nitrogen removal by anammox
bacteria. These experimental protocols aim to serve as a
useful guide that establishes a basis for standardizing
batch activity tests for use on existing, emerging and
innovative treatment processes. It was decided to start the
order of presentation with EBPR systems involving
PAOs as the processes are complex and include all three
biochemical activated sludge environments: anaerobic,
anoxic and aerobic.
Figure 2.1.1 Experimental facilities for activated sludge activity tests at UNESCO-IHE Institute for Water Education in the Netherlands (photo: UNESCO-IHE).
ACTIVATED SLUDGE ACTIVITY TESTS
9
20i306, 2007; Nielsen et al., 2010). In particular, efforts
have focused on developing a better understanding of the
actual EBPR metabolic mechanisms, to unravel the
microbial identity of the organisms involved, and to
optimize the required process configurations, all with the
aim of improving and increasing the EBPR process
efficiency and reliability.
2.2 ENHANCED BIOLOGICAL
PHOSPHORUS REMOVAL
2.2.1 Process description
Enhanced biological phosphorus removal (EBPR) can be
implemented in activated sludge wastewater treatment
systems by introducing an anaerobic stage at the start of
the wastewater treatment lines. High P-removal
efficiency, lower operational costs, lower sludge
production and the potential recovery of phosphorus have
contributed to its application and popularity (Mino et al.,
1998; Henze et al., 2008; Oehmen et al., 2007). EBPR is
performed by phosphorus (polyphosphate)-accumulating
organisms (PAOs) (Comeau et al., 1987; Mino et al.,
1998) that, by intracellular accumulation of
polyphosphate (poly-P), can remove higher quantities of
phosphorus (0.35-0.38 g P g VSS-1 of PAOs) than OHOs
(0.03 g P g VSS-1 of OHOs) (Wentzel et al., 2008). The
scientific,
microbiological
and
engineering
characteristics of the EBPR process have been the main
focus of research carried out during the last few decades
by different research groups (Wentzel et al., 1986, 1987;
Comeau et al., 1986, 1987; Smolders et al., 1994a,b;
Mino et al., 1987, 1998; Oehmen et al., 2005a, 2005c,
Anaerobic
PAOs are heterotrophic organisms. However, unlike
OHOs, PAOs have the unique capability of using
intracellularly stored poly-P to produce the required
energy (adenosine tri-phosphate, ATP) under anaerobic
conditions to store readily biodegradable organic matter
(RBCOD), such as volatile fatty acids (VFA) like acetate
(Ac) and propionate (Pr), as intracellular poly-βhydroxy-alkanoates (PHAs). Stored PHAs are later
utilized under anoxic or aerobic conditions for enhanced
phosphorus uptake, glycogen synthesis, biomass growth
and maintenance. This feature gives PAOs a competitive
advantage over other microbial populations of relevance.
Thus, PAOs can be enriched to achieve EBPR by
recycling activated sludge through alternating the
anaerobic and anoxic or aerobic stages, while directing
the influent which is usually rich in VFA to the anaerobic
stage. A schematic representation of the PAOs’
metabolism is shown in Figure 2.2.1.
Aerobic (Anoxic)
Settling
CO2
PO4
VFA
Influent
Gly
(PO4, VFA)
PP
PHA
Effluent
PO4
Gly
PHA
PP
O2 (NOx)
RAS
WAS
Liquid phase
Anaerobic
Aerobic
Settling
PO4
PO4
VFA
PO4
PAO biomass
PP
PHA
PP
GLY
GLY
PHA
PHA
PP
GLY
Figure 2.2.1 Conceptual scheme of an activated sludge wastewater treatment plant performing EBPR, illustrating the activity of PAOs (Lopez-Vazquez, 2009;
adapted from Meijer, 2004).
10
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Mino et al., 1998). Thus, the anaerobic uptake of VFA by
PAOs results in the storage of PHAs and simultaneous
hydrolysis of poly-P and glycogen. The most common
PHA polymers stored by PAOs are poly-βhydroxybutyrate (PHB), poly- -hydroxyvalerate (PHV)
and poly-β-hydroxy-2-methylvalerate (PH2MV). Their
presence and amount depends on the VFA composition
(Ac or Pr). When Ac is the most abundant VFA in the
media, PAOs store mostly PHB (up to 90 % of the stored
PHAs) (Smolders et al., 1994a), but when Pr is the
dominant VFA, then PHAs exist mostly as PHV and
PH2MV (Oehmen et al., 2007).
In the anaerobic stage, PAOs store intracellularly the
readily biodegradable organics present in the raw influent
or settled sewage (mostly VFA) as PHAs using two other
intracellularly stored polymers that take part in the
aforementioned metabolism: poly-P and glycogen (a
polymer of glucose). Poly-P is hydrolysed and utilized by
PAOs to provide the required energy (as ATP) for the
transport and storage of VFA as PHAs (Wentzel et al.,
1986), while glycogen is used to supply the required
reducing power for the conversion of VFA into PHAs as
well as to provide the additional required energy (as
ATP) (Comeau et al., 1986, 1987; Smolders et al., 1994a;
Anaerobic
Aerobic
Settling
CO2
Effluent
VFA
Influent
Gly
Gly
(VFA)
PHA
PHA
O2
RAS
WAS
GAO biomass
Liquid phase
Anaerobic
Aerobic
Settling
VFA
PO4
PHA
PO4
PO4
GLY
GLY
PHA
PHA
GLY
Figure 2.2.2 Conceptual scheme of the microbial activity of GAOs (adapted from Lopez-Vazquez, 2009).
In addition to VFA uptake, the anaerobic hydrolysis
of poly-P and glycogen also provides the energy required
by PAOs to cover their anaerobic maintenance
requirements without carbon uptake. Consequently, the
hydrolysis of poly-P leads to the release of
orthophosphate (PO4) into the bulk liquid, which is
reflected in an increase in orthophosphate concentration
in the liquid phase during the anaerobic stage (Figure
2.2.1). In addition to the uptake of VFA present in the
influent of the activated sludge system, PAOs can also
store VFA generated by fermentative organisms in the
anaerobic stage from fermentable organics present in the
influent. Once PAOs reach the aerobic stage, they utilize
the PHAs stored in the anaerobic phase as a carbon and
energy source using oxygen as the electron acceptor; the
energy from this reaction is used to take up and store a
higher amount of PO4 than the amount previously
released in the anaerobic stage (Figure 2.2.1). This results
in the aerobic uptake and removal of phosphorus from the
liquid phase. In the aerobic stage, PHAs are also used to:
(i) replenish the intracellular glycogen pool, (ii) support
biomass growth, and (iii) cover the aerobic maintenance
energy needs of PAOs (Smolders et al., 1994b). Net Premoval from wastewater is achieved through the
ACTIVATED SLUDGE ACTIVITY TESTS
wastage of activated sludge (WAS) at the end of the
aerobic phase, when the sludge contains a high poly-P
content (Figure 2.2.1). Alternatively, denitrifying
phosphorus-accumulating organisms (DPAOs) exist
which can also take up PO4 under anoxic conditions using
nitrate or nitrite as electron acceptors (Vlekke et al.,
1988; Kuba et al., 1993; Hu et al., 2002; Kerr-Jespersen
et al., 1993; Guisasola et al., 2009). Also, PAOs, being
heterotrophic organisms, are able to take up carbon
sources
under
aerobic
conditions,
releasing
orthophosphate while the carbon source is available and
removing PO4 afterwards (Guisasola et al., 2004; Ahn et
al., 2007). However, eventually PAO can lose the
competition against OHOs due to their metabolic
adaptation to permanent aerobic conditions (Pijuan et al.,
2006). For a deeper understanding of the metabolism and
factors affecting the EBPR process, the reader is referred
to materials published elsewhere (Comeau et al., 1986;
Mino et al., 1998; Oehmen et al., 2007).
The proliferation of glycogen-accumulating
organisms (GAOs) has been observed in EBPR systems
under certain conditions (e.g. when acetate or propionate
are present as the sole carbon source, when temperatures
exceed 20 °C, at pHs below 7.0, and/or at dissolved
oxygen (DO) concentrations higher than 2 mg L-1)
(Oehmen et al., 2007; Lopez-Vazquez et al., 2009a,b;
Carvalheira et al., 2014). GAOs have an apparently
similar metabolism to that of PAOs, but they rely solely
on their intracellularly-stored glycogen pools as the
source of energy and reducing equivalents that drive the
anaerobic storage of VFA as PHAs without any
contribution from poly-P (Figure 2.2.2). Their presence
is often associated with suboptimal EBPR performance
because they do not contribute to phosphorus removal,
but compete with PAOs for substrate under anaerobic
conditions leading to the deterioration of EBPR systems
(Saunders et al., 2003; Thomas et al., 2003). Therefore,
GAOs are assumed to be an undesirable population in
EBPR systems.
11
produce DO concentrations higher than 2 mg L-1 under
aerobic conditions, (iii) provide complete mixing
conditions, (iv) allow temperature control; (v) allow pH
control, and (vi) have ports for sample collection and the
addition of influent, solutions, gases and any other liquid
media or substrate used in the test (Figures 2.2.4 and
2.2.5).
Figure 2.2.3 Vintage EBPR experimental setup with the characteristic
yellowish colour of highly enriched biomass with PAOs used at Delft
University of Technology in the early 1990s for the development of the
TUDelft bio-P metabolic model (Smolders et al., 1994a, 1994b; Murnleitner
et al., 1997) and pioneering research on the impact of temperature on EBPR
(Brdjanovic et al., 1998a) (photo: Brdjanovic, 1994).
2.2.2 Experimental setup
2.2.2.1 Reactors
To assess the EBPR process performance, batch activity
tests can be carried out under anaerobic, aerobic and
anoxic conditions depending upon the parameters of
interest and nature of the study. In any case, the
bioreactor(s) used for the execution of tests must: (i)
avoid oxygen intrusion under anaerobic and anoxic
conditions, (ii) provide enough aeration capacity to
Figure 2.2.4 Temporary experimental setup used to carry out batch tests
with activated sludge at the WWTP Haarlem Waarderpolder in the
Netherlands. This was the first study (Brdjanovic et al., 2000) in which a
separate validation of the TUDelft bio-P metabolic model was carried out
using various batch tests with mixed culture biomass from a full-scale plant
(photo: Brdjanovic, 1997).
12
Anaerobic conditions
The experimental setup used for EBPR batch activity
tests must be able to create and maintain strict anaerobic
conditions. This means that no electron acceptors
(namely oxygen, nitrate or nitrite) should be available to
the biomass during the anaerobic phase.
A redox probe can be used to monitor the creation of
anaerobic conditions when the redox values are lower
than -300 mV. The lab setup should be airtight and
equipped with an off-gas exit connected to the lid of the
bioreactor. Usually, there are three undesirable sources
of oxygen: (i) the oxygen dissolved in the influent, (ii)
the residual oxygen present in the activated sludge itself,
and (iii) oxygen intrusion from the head space. To
remove the first two listed sources, N2 gas should be
sparged under mixing conditions from the bottom of the
bioreactor for 5 to 10 min prior to the beginning of the
test and during influent addition. Sparging time will
depend on the mass transfer properties of the gas-liquid
interface, which depends on a number of factors
including: the dimensions of the bioreactor, the presence
and location of baffles, dimensions and stirring speed of
mixing blades, gas diffuser configuration and flow rate,
and medium composition. To avoid oxygen intrusion, the
headspace can be flushed either by the N2 gas already
sparged at the bottom of the bioreactor to the activated
sludge or by flushing the head space for 5 to 10 min,
depending on the volume of the head space and the gas
flow rate. A N2-gas flow rate of around 30 L h-1 is
commonly used in lab-scale fermenters with an operating
volume of up to 3 L, while a lower flow rate of about 6 L
h-1 is recommended for batch reactors with working
volumes of around 0.5-1.0 L. Sparging N2 gas from the
bottom of the bioreactor is common practice and it can
be applied prior to, at the beginning, and during the
execution of the activity tests, whereas flushing of the
head space is often used during the execution of the test
to avoid oxygen intrusion from the atmosphere when
mixing the activated sludge. Combining these two
approaches is both unusual and unnecessary.
To avoid diffusion of oxygen into the bioreactor, a
unidirectional check-valve or a water-lock (containing an
oxygen scavenger, such as NaSO2) should be connected
to the off-gas line. Alternatively, if the bioreactor is
continuously sparged with N2 gas, resulting in positive
pressure inside the bioreactor and a continuous off-gas
flow, a check valve or water-lock are not essential for
ensuring anaerobic conditions. If N2 gas is unavailable or
cannot be continuously supplied due to limitations of the
equipment, the activated sludge should be gently yet
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
completely mixed at slower speeds (much lower than 300
rpm) under airtight conditions until the dissolved oxygen
(DO) concentration drops below the detection limit
(practically zero) and the redox values are lower than 300 mV. In addition, the volume of headspace should be
reduced to minimize the risk of oxygen intrusion by
filling the fermenter to the maximum working volume
and/or by reducing the surface area of the gas-liquid
interface by adding non-reactive, floating polyurethane
foam or sponge beads. Silicon rubber stoppers and seals,
plastic and aluminium foils, among other materials, are
usually used to create airtight conditions.
In addition to avoiding oxygen intrusion, the presence
and availability of other electron acceptors (such as
nitrate or nitrite) must be prevented to keep strict
anaerobic conditions throughout the test. Their
prevention is not as straightforward as the removal of
oxygen. It often requires an adequate handling of the
activated sludge sample prior to the execution of the test.
This may involve the controlled addition of nitrifying
inhibitors under non-aerated and aerated conditions, or
sludge 'washing', as explained later in Section 2.2.3.5).
Anoxic conditions
The creation of strict anaerobic conditions means that no
electron acceptors should be present inside the
bioreactor, the creation of anoxic conditions indicates
that although DO must not be present, other electron
acceptors, such as nitrite or nitrate, must be available.
This also means that the creation of anoxic conditions
must reduce or eliminate oxygen intrusion as is done for
anaerobic phases. The experimental setup should allow
the (controlled) availability (presence) of the electron
acceptors of interest. The desired electron acceptors can
be generated either in the system itself, via a preceding
nitrification step, or be externally added as nitrate - or
nitrite - solutions at defined concentrations at the start of
the batch test or during the test. If available, DO
concentrations below the detection limit together with
redox values in between -200 to 0 mV can indicate that
the required anoxic conditions have been reached.
Usually, the latter is the most common practice because
the experimental configuration can be simplified by
doing so, and there is a better control of the required
dosing time and concentration. Nevertheless, a
combination of several reactors and experimental
stages/phases can be made to incorporate a nitrification
step between the anaerobic and anoxic phases to provide
the required electron acceptors to drive the anoxic
metabolism of EBPR cultures (Kuba et al., 1993). When
external electron donors are added, the system must have
ACTIVATED SLUDGE ACTIVITY TESTS
an adequate dosing port and a way to release the resulting
extra pressure created by injecting the liquid volume.
Aerobic conditions
Most commercially available fermenters have gas
spargers usually located at the bottom of the fermenter
just below the stirring blades of the mixer. When
supplying compressed air (e.g. either from a central or a
local/portable air compressor), these arrangements can
provide a satisfactory oxygen supply leading to DO
concentrations reaching far above the limiting conditions
of the microbial processes. As a general rule of thumb,
DO concentrations of at least 2 mg L-1 are considered
adequate for most applications. For batch reactors with a
working volume of about 3 L, a compressed air flow rate
of around 60 L h-1 (1 L min-1) can usually provide the
required aeration. However, the biomass composition
and concentration, wastewater characteristics, organic
matter and intracellular PHA content (in the case of
EBPR) may increase the DO requirements. Under these
conditions, the air flow rate should be increased so as to
maintain the DO concentration above 2 mg L-1
throughout the test. Alternatively, a pure oxygen supply
can be used instead of compressed air to increase the DO
availability under specific conditions (e.g. in industrial
applications).
In more advanced applications, it may be necessary
or desirable to carry out aerobic batch activity tests at a
constant (set) DO concentration. For these applications,
a two-way DO control can be used to define a DO set
point and keep it stable throughout the aerobic batch test.
Most advanced fermenters are equipped with such a twoway control operated by at least two solenoids with an
on/off function that alternatively supplies air or N2 gas,
depending upon the actual measured DO in the liquid
phase.
Less advanced fermenters used for the execution of
aerobic batch activity tests can be equipped with a
portable air compressor that provides an adequate air
flow rate. Aquarium stones can be placed at the bottom
of the fermenters in line with the mixing/stirring system
to distribute bubbles for good oxygen transfer. As
previously discussed, the air supply should be able to
produce and sustain a bulk liquid DO concentration of at
least 2 mg L-1 within the first 10 min.
The two most common commercially available DO
probes are the membrane-type and the optical-type. Prior
to use (and preferably also after use), they should be
calibrated according to the manufacturer's or supplier's
instructions. In addition, all connections should be
13
checked. In the case of the membrane-type DO probe, the
membrane should be clean, should not have any damage,
and the probe should be properly filled with fresh
electrolytes. Moreover, no bubbles should accumulate or
be trapped on the membrane surface. The surface of
optical probes also needs to be cleaned periodically and
the head cap replaced annually.
Mixing
Mixing of the bioreactor's content must be generous to
favour a homogenous distribution of the activated sludge
(mixture of liquid phase and biomass) and wastewater as
well as other substances (e.g. orthophosphate, nitrate or
nitrite solutions). Commonly, in most 3 L fermenters, a
mixing speed of up to 500 rpm can be applied, while
slower mixing speeds of around 100 rpm are used in
larger fermenters (of 10 L and larger). Excessive mixing
can lead to floc breakage, reducing the mass transfer
resistance through the flocs and the settling properties.
On the other hand, insufficient mixing can result in dead
zones, large flocs, sludge stratification, limited diffusion
of substrates and oxygen, and, in extreme cases, in
settling. Slower mixing speeds can be used as long as the
bulk liquid is well mixed and neither stratification nor
accumulation of solids is observed. Advanced fermenters
can have an automatic control to regulate the mixing
speed in time using a vertical axis with blade propellers.
Furthermore, the use and installation of vertical blades or
baffles connected to the inner side of the fermenter's lid
or the choice of a more efficient impeller can further
improve the mixing conditions. In less sophisticated
systems, mixing can be provided by positioning the
fermenters on stirring plates and using magnetic stirrers.
As previously mentioned, to reduce the potential oxygen
intrusion in anaerobic batch tests, the stirring speed can
be reduced as long as it does not compromise the good
mixing conditions.
Temperature control
Temperature has a strong effect on the metabolism of
PAOs and their competitors (e.g. GAOs) (Brdjanovic et
al., 1997; Lopez-Vazquez et al., 2009a,b). Therefore,
adequate and stable temperature-controlled conditions
are advisable for the execution of the batch activity tests.
Advanced fermenters are usually equipped with a double
glass wall (double-jacketed reactors) and usually water
(of a temperature similar to the target temperature in the
bioreactor) is recirculated through the double wall. The
water temperature is adjusted in the controlling and
operating console of the fermenter (through internal
heaters, heat exchangers and condensers) or by using
external heating jackets or a water bath and recirculation
devices. Depending upon the desired working
14
temperature, other fluids rather than water can be
recommended (e.g. anti-freeze solutions for temperatures
lower than 5 ºC or oils for temperatures higher than 30
ºC). Advanced systems can automatically measure the
bulk water temperature inside the bioreactor and adjust
accordingly to keep a stable temperature. However, it is
important to keep in mind that the temperature of the
cooling/heating fluid is often different (by a couple of
degrees Celsius) from the actual temperature measured in
the activated sludge. These differences occurr due to
thermal exchange efficiency and thermal exchange
between the recirculated fluid and the air during its
transport from the controller to the fermenter, in
particular when the operating temperature is significantly
different from the ambient (room) temperature. Under
such conditions, it is recommended to adjust the fluid
temperature in the controller unit until the target
temperature in the liquid phase is reached and remains
stable.
Besides the individual temperature control that a
fermenter may have, the entire experimental setup can be
located inside a temperature-controlled room set at the
target temperature. Nevertheless, if temperature is not an
issue and the batch tests can be performed at room or
ambient temperature in a defined location, there should
be the certainty that the temperature will not fluctuate
considerably (less than ± 1-2 ºC) from the preparation
until the end of the test. In any case, the temperature of
execution must be always recorded and reported.
Last but not least, it is important to mention that both
the activated sludge and the wastewater or synthetic
medium of the study (whenever applicable) must have
the same temperature prior to the execution of the activity
tests to avoid temperature shocks and fluctuations that
may compromise the outcomes of the tests. Under these
circumstances, all the activated sludge, wastewater and
solutions need to be exposed to the working temperature
and their actual temperature must be monitored until they
reach the target temperature. The temperature adjustment
of the activated sludge must be carried out with no
electron donor present (i.e. no external carbon source).
Usually, only a short exposure time of the biomass to the
desired temperature of maximum 1 to 2 h is necessary.
Whenever needed, the samples could also be
acclimatized for longer periods (up to 3-4 h) until the
target temperature is reached, but special attention must
be paid to avoid compromising the metabolic activity of
PAOs (e.g. leading to the consumption of the intracellular
compounds). If the temperature difference between the
activated sludge and substrate media or wastewater is
high or if the temperature effects are of interest (e.g.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
higher than 5 °C to assess a potential temperature shock),
the tests must be conducted as soon as the target
temperature is reached.
Usually, most of the tests are executed around 20 °C,
but it can be as low as 5 ºC (to assess the biomass activity
under winter/cold climate conditions) (Brdjanovic et al.,
1997) or as high as 30 - 35 °C for tropical conditions or
industrial applications (Cao et al., 2009; Ong et al.,
2014), and even up to 55 °C for thermophilic conditions
(Lopez-Vazquez et al., 2014). Tests are rarely performed
below 5 °C because in practice the temperature of
municipal wastewater is seldom colder and is usually
around 7 - 12 °C.
pH control
pH is an important operating parameter for EBPR (and
many other) processes. This is particularly because the
metabolism of PAOs during the anaerobic uptake of
carbon (VFA) will result in higher P-release levels at
higher pH and lower P release at lower pH (Smolders et
al., 1994a). Also, mixing and the vigorous sparging of N2
gas or compressed air can strip the dissolved CO2 out of
the solution and raise the pH above 7.0. (e.g. in the range
of 7.8-8.5), affecting a number of biological and physical
processes.
On the other hand, the biological removal of
constituents, such as phosphate by PAOs, tends to
decrease the buffering capacity of the liquid and change
the pH during an experiment. The alkalinity of the
wastewater or other solutions added can increase the
buffering capacity of the bulk liquid and reduce the
fluctuations. CO2 sparging can compensate for the CO2
that strips out when mixing or sparging compressed air
or N2 gas. Thus, similar to temperature, pH must be stable
prior to, throughout and until the end of the EBPR batch
test.
Under certain circumstances, different pH set points
can be applied during different experimental phases (e.g.
a pH of 7.5 under anaerobic conditions followed by a pH
of 7.0 in the aerobic stage). An acceptable fluctuation pH
range is assumed to be ± 0.1. In this regard, the use of a
two-way pH controller (for acid and base addition, such
as HCl and NaOH, respectively) is recommended.
Advanced bioreactor systems usually have pH
control settings, but simpler yet reliable external pH
controllers can also be used. In less advanced systems,
pH levels can be controlled through the manual addition
of acid and base solution. The usual molarity of acid and
base solutions for pH control is around 0.2 to 0.4 M. If
ACTIVATED SLUDGE ACTIVITY TESTS
manual pH control is applied, the concentrations can be
lower (e.g. 0.1 M). Depending on the activity of the
sludge, different molarities can be used. If the molarity
of the solutions is too high, it may lead to sudden pH
changes, where the pH values may drop or increase
drastically around the set point, crossing the lower or
upper limit of the pH control settings and even oscillating
below and above the pH set point. Lower molarities may
lead to a slow response to adjust the pH to the desired pH
set point, which in extreme cases might not be reached
and create a considerable dilution of the activated sludge
in the bioreactor.
Due to the fast initial speed of microbial conversions,
the potential acid or base consumption will be higher at
the beginning of the tests or when switching from one
phase to another (e.g. from anaerobic to aerobic), but it
will usually stabilize by the end of the test. In any case, a
considerable deviation from the pH set point (e.g. of
more than ± 0.10) must be corrected, preferably within 510 sec. The actual pH measured in the liquid phase during
the experiment should always be reported.
Certain pH controllers have specific settings that
should be adjusted to maintain a stable pH, such as the
volume (stroke) of the pulses of acid and base addition
and the response time in between acid and base addition
pulses. The time in between the acid or base addition
pulses should be adjusted to the time that is needed for
the system to obtain homogenously mixed conditions
after the addition of acid or base.
Similarly to temperature, if pH shocks are to be
studied, the EBPR batch activity tests must start as fast
as the pH of the activated sludge reaches the target pH of
the study. Any required pH adjustment must be
performed preferably in less than 5 min prior to the start
of the test to avoid any premature or side effects on the
metabolism of PAOs (e.g. leading to certain P release or
consumption of polymers stored intracellularly). The use
of sulphuric acid and alkaline, phosphate buffers and
Tris(hydroxymethyl)aminomethane (Tris)
solutions
must be avoided. This is mostly since they can lead to
interferences such as benefiting sulphate-reducing
bacteria (SRB) over PAOs (Saad et al., 2013, RubioRincon et al., 2016, submitted), enhancing P
precipitation with carbonate species (chemical Pprecipitation) (Barat et al., 2008) or increasing the
salinity levels beyond those that PAOs can withstand
(Welles et al., 2014). It is needless to say that proper pH
control is essential for the success of experiments as even
very short exposure of biomass to extreme pH (low or
high) will quite certainly affect the biomass irreversibly.
15
All pH meters and sensors should be calibrated
immediately (and preferably checked afterwards)
according to the manufacturer's or supplier's instructions
and all the connections should be checked. Special
attention must be paid to the selection and use of pH
sensors that can stand the particular characteristics of the
wastewater and EBPR sludge subject to study. For
instance, high salinity, high chlorides or high H2S
concentrations can lead to interferences if the pH sensors
cannot tolerate the higher concentrations. The reader
should always verify in manuals, booklets and/or with
providers and suppliers if the pH sensors and meters to
be used are suitable for the particular wastewater
characteristics to be tested.
Sampling and dosing ports
The reactors/fermenters used to carry out the batch
activity tests should also have conveniently located
sampling and dosing ports to ensure the collection of
representative samples from the liquid phase as well as
favour a fast dispersion or mixture of any substance or
solution added to the bulk of the liquid. The sampling
ports can be composed of flexible (rubber or plastic)
tubing with an inner diameter that makes it possible to
connect different syringes of 5, 10 or 20 mL volume, but
also of smaller or bigger volumes (e.g. of 1 or even 50
mL). The sampling port inside the fermenter needs to
reach a favourable depth and location to allow a
satisfactory sampling before, during and at the end of the
test. Usually, the sampling port can be located at the
middle level of the lowest third or quarter of the
bioreactor's working volume subject to the provision of
well-mixing conditions. The most important requirement
is to obtain a representative sample from a well-mixed
bioreactor.
Regarding the dosing ports, they need to be located
and positioned in such a way that they allow a fast
dispersion of the solutions or substances added into the
bioreactor. They can be well defined injection ports
located on the lid of the bioreactor or flexible openings
(e.g. through the septum). A formal and structured
location and integration of the sampling ports into the
fermenter configuration is mostly required when working
with airtight reactors to avoid oxygen intrusion.
Moreover, the use of lab clips (or similar devices) is
recommended to close the tubing of the sampling ports
or temporarily unused dosing ports and avoid spillages
and splashes caused by a possible increase in the internal
pressure in the bioreactor. To counteract any potential
under or overpressure, a needle can be inserted into a
septum located on the lid of the bioreactor and a tedlar
bag (or a similar flexible container filled with an inter gas
16
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
- e.g. nitrogen) can be connected to the needle to connect
the gas phase of the headspace with the inert gas.
equipped with the required experimental and analytical
equipment (Figure 2.2.4).
2.2.2.2 Activated sludge sample collection
If the batch activity tests cannot be performed in situ
on the same day, an activated sludge sample can be
collected at the end of the aerobic stage. Afterwards, the
sampling bucket can be properly stored and transported
in a fridge or in ice (below or close to 4 °C) under nonaerated conditions and the activity tests should be
performed not later than 24 h after sampling. The
sampling collection at the end of the aerobic stage and
storage under non-aerated conditions at the lower
temperature can help to preserve the original biomass
condition by slowing down the bacterial metabolism.
Therefore, in principle it is advised not to aerate the
activated sludge samples since this could lead to P release
and oxidation of intracellular compounds (such as PHAs,
glycogen and even poly-P). This also implies that the
biomass present in the activated sludge sample needs to
be ‘re-activated’ and acclimatized to the target pH and
temperature of interest prior to the execution of the batch
activity tests. In any case, the in situ execution of the
batch activity tests is preferable. The total volume of
activated sludge to be collected depends on the number
of tests, bioreactor volume and total volume of samples
to be collected to assess the biomass activity. Often 1020 L of activated sludge from full-scale wastewater
treatment plants can be collected. On the other hand,
samples collected from lab-scale reactors rarely contain
more than 1 L because lab-scale systems are usually
smaller (from 0.5 to 2.2 L and under certain cases up to
8-10 L) and the maximum volume that can be withdrawn
from lab-scale reactors is often set by the daily
withdrawal of the excess of sludge from the system
(which is directly related to the applied sludge retention
time (SRT) and, consequently, defined by the growth rate
of the organisms).
Contrary to some wastewater treatment processes, the
sampling time and location of an EBPR activated sludge
sample is highly dependent on the type of batch activity
test to be conducted. The latter is based on the alternating
anaerobic-(anoxic)-aerobic conditions required by the
physiology of PAOs. Thus, a fresh sample should
preferably be collected at the end of the preceding
reaction stage. Thereby, for an anaerobic batch test, the
activated sludge sample should be collected at the end of
the aerobic phase at the full- or pilot-scale wastewater
treatment plant or 'parent' laboratory bioreactor, whereas,
for an aerobic batch test, the sample can be collected at
the end of the anaerobic or anoxic phase depending on
the system configuration. For the execution of anoxic
tests, the sludge can be collected at the end of the
anaerobic stage. Alternatively, samples collected in the
aerobic phase can be used to execute sequential
anaerobic-aerobic, anaerobic-anoxic or anaerobicanoxic-aerobic batch tests.
Certainly, the sampling location will depend on the
system configuration. In full- and pilot-scale wastewater
treatment plants, the physical borders between stages
must be identified prior to sampling. In extreme cases,
where the phases are not (physically) well defined, the
redox limits or boundaries need to be determined with the
use of a DO meter, redox meter and/or by determination
of the nitrate and nitrite concentrations. In lab-scale
systems (usually operated on a time-base mode), the
sample collection can be relatively easier, since the
reaction time defines the length of the stages. To obtain
homogenous and representative samples, the sludge
samples must be collected in sampling spots where wellmixed conditions take place. Ideally, batch activity tests
must be performed as soon as possible after collection (in
less than 1-2 h for tests to be conducted with sludge
collected at the end of an aeration tank/phase or in a few
minutes (2-3 min) for sludge collected at the end of an
anaerobic or anoxic phase). In lab-scale systems, in
principle, this should not be a problem if the batch
activity tests are performed in the same laboratory and
their execution is coordinated and synchronized with the
operation cycle of the lab bioreactor. Also, at full- and
pilot-scale treatment plants, batch activity tests can be
performed in situ shortly after the collection of activated
sludge if the sewage plant laboratory is conditioned and
2.2.2.3 Activated sludge sample preparation
For batch activity tests performed in situ, in principle, the
sludge will be merely transferred from the parent
bioreactor (in the case of a lab-enriched sludge) or
reaction tank (in the case of pilot-scale or full-scale
plants) to the fermenter or bioreactor where the activity
tests will take place. Usually, the sludge transfer must
take place before the end of the reaction phase that
precedes the reaction phase of interest. Then, the batch
activity test can start as soon as the desired pH, redox
conditions and temperature are adjusted. During the
adjustment until the test starts, the same or similar
conditions to those prevailing when the sludge samples
ACTIVATED SLUDGE ACTIVITY TESTS
were collected should be kept. This means that sludge
samples collected at the end of the anaerobic stage should
be kept under anaerobic conditions and therefore must
not be aerated or exposed to the presence of any electron
acceptor. Similarly, samples collected in the aerobic
stage must be aerated and sludge samples collected in the
anoxic stage should not be aerated. If desirable, a few
milligrams of nitrate can be added to activated sludge
samples collected in the anoxic tanks (to a final
concentration of ~5 mg NO3-N L-1) to maintain anoxic
conditions as long as it is necessary.
If only EBPR tests will be executed and nitrification
tests are not of interest, then a nitrification inhibitor can
be added to the sludge sample immediately after the
sludge is transferred to the fermenter (e.g. Allyl-Nthiourea: ATU to a final concentration of 20 mg L-1). This
will restrain nitrification and consequently (i) avoid
higher oxygen consumption in aerobic EBPR batch tests
and, (ii) limit the accumulation of nitrate (or nitrite) if
samples are aerated prior to the execution of anaerobic
tests. Should the real and actual conditions be assessed,
then the corresponding batch activity tests must be
conducted right away after sludge collection with the
minimum adjustments and stable conditions required
(e.g. for pH and temperature). A comprehensive
sampling procedure must be carried out before, during,
and after the tests to document the results obtained.
However, in addition, the execution of batch activity tests
under favourable conditions to PAOs is always
recommended. This can help to (i) assess the EBPR
potential that the system can have, (ii) benchmark the
EBPR plant activity, (ii) detect interferences, and (iv)
contribute to the definition of improvement strategies.
As described elsewhere, interferences to PAOs can be
(but are not limited to) the presence of nitrate or nitrite in
the aerobic sludge samples collected to execute anaerobic
batch tests, the existence of RBCOD in anaerobic
samples for the performance of aerobic or anoxic tests, or
the detection of nitrite in anoxic samples intended to
carry out aerobic batch tests. Thus, if tests are designed
to be conducted under favourable conditions to PAOs
then such interferences should be avoided. After this,
sludge samples can be exposed shortly (for 1 or
maximum 2 h) to a pre-treatment or preparation step as a
troubleshooting strategy. For instance, to remove the
nitrate present in an aerobic sample intended for an
anaerobic batch test (~5-10 mg NO3-N L-1), after
collection the sludge can be transferred to the airtight
batch bioreactor and gently mixed under non-aerated
conditions. The nitrate (and nitrite) concentration can be
monitored until it drops below the detection limits. Rapid
17
detection techniques, such as nitrate and/or nitrite
detection paper strips (e.g. Sigma-Aldrich), can be rather
useful here. Once nitrate is no longer observed, the
corresponding anaerobic batch test can start. If RBCOD
is detected in a sample taken from an anaerobic tank, the
anaerobic conditions can be extended after the sludge is
transferred to the airtight bioreactor until no more
RBCOD is observed, before the anoxic or aerobic test
starts. Anoxic samples to carry out aerobic tests where
nitrate is observed do not need pre-treatment since nitrate
is innocuous to PAOs under aerobic conditions.
However, if nitrite is detected, it must be removed
because it has been proven to be rather inhibitory and
even toxic to PAOs under certain aerobic conditions
(Pijuan et al., 2010; Zhou et al., 2012; Yoshida et al.,
2006; Saito et al., 2004; Zeng et al., 2014). To avoid
nitrite, a similar approach like the one previously
described for the removal of nitrate can be applied.
When the batch tests cannot be performed in situ and
sludge samples are stored under cold conditions (at
around 4 °C), sludge samples need to be 're-activated'
because the cold temperature considerably slows down
the bacterial metabolism. Due to the particular
physiology of PAOs, to reactivate the sludge it must be
aerated for 1-2 h at the pH and, particularly, at the
temperature of execution of the batch activity tests. This
procedure will help to remove residual biodegradable
organics. If only EBPR activity tests will be carried out,
a nitrification inhibitor should be added (e.g. ATU at a
final concentration in the bioreactor of 20 mg L-1) prior
to aeration. If a potential interference is detected (e.g.
nitrate, nitrite or RBCOD) prior to the 1-2 h aeration
period, the sludge reactivation can start with a pretreatment step at the temperature and pH of interest.
Afterwards, even if the objective is only to assess the
anoxic or aerobic activity of PAOs, the EBPR activity
tests should start with an anaerobic incubation stage
using synthetic or real wastewater as a feed. This practice
will ensure that the PAOs will have PHAs intracellularly
stored to carry out their aerobic or anoxic metabolisms.
Nevertheless, in general, the objective of the
experimental plan should be to minimize as much as
possible the needs for transport, cooling, storage and
reactivation of the sludge (among other potential steps).
Whenever possible, it is best to always use 'fresh' sludge
(and substrate/media).
2.2.2.4 Substrate
When real wastewater (either raw or settled) is used for
the execution of activity tests, it can be fed in a relatively
18
straightforward manner to the bioreactor/fermenter. For
normal (regular) conditions, the feeding step takes place
at the beginning of the anaerobic stage to favour the VFA
uptake by PAOs and the intracellular availability of
PHAs. If judged necessary, a rough filtration step (using
10 μm pore size filters) can be used to remove the
remaining debris and large particles present in the raw
wastewater. If the activity tests need to be performed at
different biomass concentrations, the treated effluent
from the plant can be collected and used for dilution
(assuming that solids effluent concentrations are
relatively low, e.g. 20-30 mg TSS L-1). If different carbon
or phosphorus sources and concentrations are to be
studied, the plant effluent can also be used to prepare a
semi-synthetic
media
containing
a
RBCOD
concentration of between 50 and 100 mg COD L-1.
Batch activity tests are frequently performed with
synthetic wastewater: (i) to ensure a better control of the
experimental conditions, (ii) to create the desired redox
conditions, (iii) to study and assess the effects of different
wastewater composition, or (iv) to evaluate the inhibitory
or toxic effects of certain solutions or compounds.
However, such practice can be expensive due to the
potentially large amount of chemicals needed.
Depending upon the nature, purpose and sequence of
the activity tests (anaerobic, anoxic or aerobic), the
carbon and phosphorus concentrations present in a
synthetic wastewater can vary since they are usually the
subject of removal and study. Moreover, the
concentrations may be adjusted proportionally to the
length or duration of the test. Usually, concentrations of
up to 50 and 100 mg RBCOD L-1 are used when activated
sludge samples are obtained from full-scale plants and of
up to 400 mg L-1 in the case of lab-enriched cultures
(though higher concentrations are sometimes applied).
Regarding the phosphorus concentrations, synthetic
solutions can contain none or low P concentrations when
assessing the anaerobic P release or up to 100-120 mg
PO4-P L-1 when testing the maximum aerobic P-uptake
activities. However, regardless of the nature of the
activity test, synthetic wastewater must contain the
required macro- and micro-elements (especially
potassium and magnesium but also calcium, iron, zinc,
cobalt, among others) in sufficient amounts and suitable
species that PAOs require in order to avoid any metabolic
limitation that may jeopardize the outcomes of the batch
activity tests (Brdjanovic et al., 1996). A suggested
synthetic wastewater recipe for an initial orthophosphate
concentration of 20 mg P L-1 can contain per litre
(Smolders et al., 1994a): 107 mg NH4Cl, 90 mg
MgSO4·7H2O, 14 mg CaCl2·2H2O, 36 mg KCl, 1 mg
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
yeast extract and 0.3 mL of a trace element solution (that
includes per litre 10 g EDTA, 1.5 g FeCl3·6H2O, 0.15 g
H3BO3, 0.03 g CuSO4·5H2O, 0.12 g MnCl2·4H2O, 0.06 g
Na2MoO4·2H2O, 0.12 g ZnSO4·7H2O, 0.18 g KI and 0.15
g CoCl·6H2O). Overall, it is important to underline that
the minimal concentrations of K and Mg need to be
proportional to the phosphorus concentration following a
molar 1:1:3 Mg:K:P ratio. This is mostly because Mg and
K are essential for poly-P formation since they serve as
counter-ions in poly-P. If desired, the synthetic
wastewater can be concentrated, sterilized in an
autoclave (for 1 h at 110 ºC) and used as a stock solution
if several tests will be performed in a defined period of
time. However, the solution must be discarded if any
precipitation or loss of transparency is observed.
For experiments performed with lab-enriched
cultures, it is best to execute the tests with the same
(synthetic) wastewater used for the cultivation according
to the carbon or phosphorus concentrations of the study.
Alternatively and similar to full-scale samples, the
effluent from the bioreactor can be collected, filtered
through rough pore size filters to remove any particles,
and used to prepare the required media with the desired
carbon and phosphorus concentrations for the execution
of the activity tests.
For the execution of (conventional) anoxic tests,
nitrate and nitrite solutions can be prepared to create the
required anoxic redox conditions. For this purpose,
different stock solutions can be prepared using nitrate
salts and nitrite salts. However, for practical applications,
it is recommended to follow a step-wise approach and
carefully monitor their addition so their concentrations in
the bulk water do not exceed more than 10 mg L-1 in the
case of nitrite and 20 mg L-1 for nitrate. This will avoid a
potential inhibitory effect due to nitrate or nitrite
accumulation as described elsewhere (Saito et al., 2004;
Yoshida et al., 2006; Pijuan et al., 2010, Zhou et al.,
2007, 2012). Moreover, pH levels lower than 7.0, in
combination with higher nitrite concentrations, can be
comparatively more inhibiting to PAOs because nitrite
can be present as free nitrous acid (FNA) - the
'protonated' species of nitrite. At pH 7.0, Zhou et al.
(2007) and Pijuan et al. (2010) observed 50 % inhibition
of the anoxic and aerobic metabolism of PAOs at FNA
concentrations of 0.01 and 0.0005 mg HNO2-N L-1,
respectively (equivalent to 45 and 2 mg NO2-N L-1 at pH
7.0). Therefore to ensure their availability during the
execution of the anoxic tests, the nitrite or nitrate
concentrations need to be monitored during the tests and
depending on their concentrations, nitrite or nitrate
solutions will need to be added in different steps.
ACTIVATED SLUDGE ACTIVITY TESTS
When tests are conducted to assess the potential
inhibitory or toxic effects of given compounds at
different concentrations, concentrated stock solutions
can be prepared and added during the test at the
concentrations of interest. Tests performed to assess
whether the inhibitory or toxic effects are reversible must
be carried out after 'washing' the biomass to remove the
inhibiting or toxic compound(s). The washing step is
often performed by consecutive settling and resuspension of the sludge sample in a carbon-free media
(either fully synthetic or using a treated effluent after
filtration) under the redox conditions of interest.
Similarly, when maintenance tests need to be executed, a
carbon- and phosphorus-free media can be used at the
redox conditions and operating conditions of the study.
2.2.2.5 Analytical procedures
Most of the analytical procedures required (for the
determination of total P, PO4, NH4, NO2, NO3, MLSS,
MLVSS, among others) should be performed following
standardized and commonly applied analytical protocols
detailed in Standard Methods (APHA et al., 2012). VFA
determination (for acetate, propionate and even other
volatile fatty acids) can be conducted by gas
chromatography
(GC),
high
pressure
liquid
chromatography (HPLC) or by applying regular
analytical determination protocols. However, contrary to
most of the analytical parameters of interest, the
determination of PHAs and glycogen requires a more
demanding sample preparation and sophisticated
equipment and procedures, and in addition, their
determination procedures are only to be found in
specialized scientific literature. Therefore, the analytical
procedures for the determination of PHAs and glycogen
are described in more detail in the following paragraphs
in this chapter.
• PHA
As mentioned earlier, the most common PHA polymers
stored by PAOs are poly-β-hydroxy-butyrate (PHB),
poly-β-hydroxy-valerate (PHV) and poly-β-hydroxy-2methyl-valerate (PH2MV). Their relative presence and
stored amount depends on the VFA composition (Ac or
Pr) and the type of metabolism involved in the storage
(PAO or GAO metabolism). For enriched lab cultures
performing EBPR (where PAOs are the dominant
organisms composed of more than 90 % of the total
biomass), and when Ac is the most abundant VFA, PAOs
store VFA mostly as PHB (up to 90 % of PHAs)
(Smolders et al., 1994a). However, when Pr is the
dominant VFA, then PHV and PH2MV can be composed
19
of up to 45 % and 53 % of the total PHAs stored,
respectively (Oehmen et al., 2005c). When GAOs are
present in EBPR systems, the sludge stores higher
amounts of PHV too. For instance, lab-enriched GAO
cultures (comprising more than 90 % of the total
biomass) cultivated with Ac as VFA leads to a PHB and
PHV accumulation of around 73 % and 26 %,
respectively (Zeng et al., 2003a, Lopez-Vazquez et al.,
2007, 2009a), whereas an enriched PAO culture
cultivated under similar conditions contains mostly PHB
and less than 10 % PHV (Smolders et al., 1994a).
Meanwhile, GAO lab-systems fed with Pr result in
practically no PHB accumulation, but up to 43 % PHV
and 54 % PH2MV (Oehmen et al., 2006). It is important
to underline that the PHA analytical determination
technique provides the PHA contents of the MLSS quite
accurately. This implies that a precise determination of
MLSS is equally important to obtain correct PHA
concentrations and to accurately determine net
conversions during a biochemical stage.
Compared with full-scale EBPR systems, the
determination of PHAs in lab-scale enriched EBPR
systems is usually easier since lab-scale systems are
smaller and, more importantly, EBPR cultures are
enriched with PAOs (> 90 %) (Oehmen et al., 2004,
2006; Lopez-Vazquez et al., 2007) and consequently, the
intracellular PHA contents can reach up to 10 % of the
total MLSS concentration depending upon the VFA type
available (Lopez-Vazquez et al., 2009a). On the other
hand, in the best case, the PHA contents accumulated in
the mixed biomass from full-scale systems reach between
1 and 2 % of the total MLSS concentration because PAOs
(and GAOs) hardly comprise more than 15 % of the total
bacterial population (Lopez-Vazquez et al., 2008a). This
implies that the analytical determination of PHAs from
full-scale samples may not always be suitable, reliable or
therefore representative of the direct collection of grab or
composite samples. In extreme cases, PHA contents may
fall below the detection limit. From an economic
perspective and in view of the required resources (in
terms of analytical equipment, costs of chemical
consumables and highly qualified lab staff), the PHAs
determination will probably not be (cost) effective when
performed on samples from a full-scale plant.
Alternatively, to assess the potential accumulation of
PHAs in full-scale systems, real full-scale sludge samples
can be used to execute batch activity tests under more
favourable and controlled conditions for EBPR that can
maximize accumulation of PHAs and facilitate its
analytical determination (Lanham et al., 2014).
Nevertheless, this latter approach still requires the
20
analytical determination of PHAs to be performed with
high precision.
Regarding the analytical determination of the
different PHA polymers, it has been a matter of
discussion and improvement since the late 1990s
(Baetens et al., 2002). So far, the most reliable method
involves two slightly different procedures (Oehmen et
al., 2005b): (i) one for the determination of PHB and
PHV polymers, and (ii) another for the determination of
PHV and PH2MV.
For both determination procedures, activated sludge
samples must be collected in situ in (15 mL)
centrifugation tubes. The sample volume should be
sufficient to obtain around 20 mg of TSS. To preserve the
sample, 4-5 drops of paraformaldehyde (37 %
concentration) have to be added to the plastic
centrifugation tube (in a fume hood) prior to collection of
the sample, and once the sample is taken, it should be
stored temporarily at 0-4 °C for around 2 h. To remove
the remaining paraformaldehyde and dissolved solids in
the liquid phase, samples need to be washed twice with
tap water. The washing steps include: (i) centrifugation
(for 10 min at 4,500 rpm), (ii) careful withdrawal of the
supernatant by decanting (if a solid pellet is formed) or
otherwise with a pipette, avoiding the removal of any
particle or solid, (iii) tap water addition (10 mL), and (iv)
re-suspension with a vortex. After the second washing
step, the sample must be centrifuged one more time, and
supernatant must be discarded. Afterwards, the sample
must be stored at -20 °C and subsequently freeze-dried
in a lyophilizer at -80 °C and 0.1 mbar for 48 h (or
longer), until the sample is fully dried. Once the sample
has been freeze-dried, the digestion, esterification and
extraction procedures can start.
As described by Oehmen et al. (2005b), for PHB and
PHV determination, 20 mg of the freeze-dried sample
can be transferred to a digestion tube and added to 2 mL
of an acidified methanol solution containing a 3 %
sulphuric acid (H2SO4) concentration and approximately
100 mg L-1 of sodium benzoate. Afterwards, samples are
digested and esterified for 2 h at 100 °C. After digestion
and esterification, samples are cooled down to room
temperature; distilled water is added and mixed
vigorously. 1 h of settling time must be provided to
achieve a phase separation. The chloroform phase can be
transferred to a vial, dried with 0.5-1.0 g of granular
sodium sulphate pellets and separated from the solid
phase. Standard solutions can be prepared in parallel at
defined concentrations using commercial co-polymers of
R-3-hydroxybutyric acid (3HB) and R-3-hydroxyvaleric
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
acid (3HV) copolymer (7:3). After extraction and
esterification, 3 μL of the liquid phase can be injected
into a chromatograph. Certain recommended
characteristics and operating conditions of the
chromatograph are: (i) to be equipped with a DB-5
column (30 m length × 0.25 mm I.D. × 0.25 μm film), (ii)
to apply a 1:15 split injection ratio, (iii) to use helium
(He) as the carrier gas at a flow rate of 1.5 mL min-1, (iv)
to be equipped with a flame ionization detector (FID)
operated at 300 °C with an injection port at 250 °C, and,
(v) to vary the oven temperature starting at 80 °C for 1
min, increasing 10 °C min-1 up to 120 °C, and then to
further increase it at a temperature pace of 45 °C min-1
up to 270 °C , and hold it at 270 °C for 3 min. When
following this procedure and conditions, the PHB and
PHV peaks will show up around 2 and 3 min after
injection.
Alternatively, for PHB and PHV determination,
another procedure followed by Smolders et al. (1994a)
involves the addition of 20 mg of freeze-dried biomass to
1.5 mL of dichloroethane, and 1.5 mL of concentrated
HCl as well as 1-propanol 1:4 (in volume). 1 mg of
benzoic acid in the 1-propanol solution is added as
internal standard. Samples are digested and esterified for
2 h at 100 °C and, at least every 30 min, samples are
vortexed. After cooling, 3 mL of distilled water is added.
The contents are mixed vigorously on a vortex and
afterwards centrifuged for a few minutes to obtain a
satisfactory and well-defined phase separation. About 1
mL of the lower (organic) phase is drawn off and filtered
over a small column of dried water-free sodium sulphate
into GC sample vials. As a recommendation, 3 standards
must be run for every series of 15 samples. When using
this method, 1 μL of the lower liquid phase from the
solution can be injected into a gas chromatograph
equipped and operated as follows: (i) using a HP
Innowax column (30 m length × 0.32 mm I.D. × 0.25 μm
film), (ii) applying a 1:10 split injection ratio, (iii) using
He as the carrier gas (at a flow rate of 6.3 mL min-1), (iv)
operating a FID at 250 °C, applying an injection
temperature of 200 °C, (v) with an initial oven
temperature of 80 °C kept for 1 min, that increases to
130 °C at temperature pace of 25 °C min-1, and then to
210 °C at 15 °C min-1 and finally held at 210 °C for 12
min. This long final time is recommended to elute
propylesters of no interest (e.g. from cell wall
constituents). Last but not least, the PHA contents of the
biomass is reported as a percentage of the MLSS
concentrations, which is used to calculate the PHA
concentrations.
ACTIVATED SLUDGE ACTIVITY TESTS
If the activated sludge samples contain high
concentrations of salts, a saline washing solution with a
similar osmotic strength like that of the original sample
should be used instead of tap water. This will avoid the
cytolysis of the cells and preserve the intracellular
compounds (such as PHAs and glycogen) avoiding their
dissolution and potential loss through the supernatant.
However, when a saline washing solution is used, the
high concentration of total dissolved solids (TDS) in the
remaining liquid (after centrifugation) may precipitate
and lead to apparent deviations in the MLSS
concentrations of the original sample, from which the
PHA contents will be determined. To compensate, a
correction factor will be needed to take into account the
potential effect of the TDS on the final solids sample
when the PHA concentrations are determined.
Although the two previous analytical procedures can
be rather accurate for the determination of PHB and
PHV, none of them can, without any further
modification, be satisfactorily used for the determination
of PH2MV (of particular importance when propionate is
present as a carbon source in EBPR cultures). Thus, to
improve the PH2MV extraction, Oehmen et al. (2005b)
recommend applying the same procedure described for
PHB and PHV determination, but using an acidified
methanol solution containing 10 % H2SO4 (instead of 3
% H2SO4) and extending the digestion phase at 100 °C
to 20 h. Since a commercial product to be used as a direct
standard for PH2MV determination is not available,
Oehmen et al. (2005b) recommended the use of 2hydroxycaproic acid which is assumed to have a similar
relative response to that of PH2MV (based on the fact that
these two molecules are isomers of each other). This
procedure has proven useful for the simultaneous
determination of PHV and PH2MV, but not for PHB. As
a consequence, if the three polymers (PHB, PHV and
PH2MV) must be determined, the two different
determination procedures must be performed. Further
details about the analytical PHA determination
techniques can be found in the original sources (Baetens
et al., 2002; Oehmen et al., 2005b). From a microscopic
visualization perspective, Nile blue A stain can be used
to qualitatively visualize PHAs and Neisser stain for
poly-P (Mino et al., 1998; Mesquita et al., 2013). Further
details about the microscopic observation of these and
other intracellular polymers and the use of different stains
can be found in Chapter 7 on Microscopy.
• Glycogen
EBPR cultures utilize glycogen as a source of energy and
reducing power for the storage of PHAs. Glycogen
21
(C6H10O5) is a multi-branched polysaccharide of glucose
(C6H12O6) similar to starch and cellulose but with a
different glycosidic bond and geometry between
molecules (Dircks et al., 2001; Wentzel et al., 2008). Its
relative presence and intracellular storage by EBPR
cultures depends on the VFA composition (Ac or Pr),
influent P/C ratio, and dominant organisms (either PAOs
or GAOs) (Schuler and Jenkins, 2003; Oehmen et al.,
2007). In enriched lab PAO cultures cultivated with Ac
as carbon source (where PAOs compose of more than 90
% of total biomass), at influent P/C ratio lower than 0.04
mol mol-1, the glycogen fractions can reach up to 20 % of
the total MLVSS concentration, whereas at influent P/C
ratios higher than 0.04, the glycogen fractions are usually
lower than 15 % (Smolders et al., 1995; Schuler and
Jenkins, 2003; Welles et al., 2016, submitted). Similarly,
lab-enriched PAOs cultures cultivated with Pr as carbon
source tend to store less intracellular glycogen that often
does not reach more than 15 % MLVSS since PAOs'
anaerobic metabolism on Pr requires less glycogen
hydrolysis for anaerobic P release and intracellular PHA
storage (Oehmen et al., 2005c). Conversely, GAO
cultures enriched in the laboratory using Ac or Pr as
carbon source can have glycogen fractions as high as 30
% MLVSS regardless of the carbon source fed (Filipe et
al., 2001b; Zeng et al., 2003a; Oehmen et al., 2005a,c;
Dai et al., 2007; Lopez-Vazquez et al., 2009a). Similar
to PHA determination, the determination of the
intracellular glycogen content may be easier to estimate
in lab-scale systems (where EBPR cultures can comprise
more than 90 % of the total microbial population) but it
will not be so straightforward in full-scale systems since
PAOs (and GAOs) hardly comprise more than 15 % of
the total bacterial population (Lopez-Vazquez et al.,
2008a). Consequently, the glycogen fractions present in
full-scale EBPR systems may hardly reach more than 5
% of the total MLVSS concentrations, which makes its
determination more difficult and challenging when
compared to lab-scale systems. Nevertheless, it may be
still feasible but it requires analytical determination of
high precision (Lanham et al., 2014). Glycogen
(C6H10O5) is a multi-branched polysaccharide; it should
be hydrolysed and extracted prior to its determination.
Thus, different methods have been proposed for the
analytical determination of glycogen, ranging from
enzymatic hydrolysis tests (Parrou and Francois, 1997) to
biochemically-based (Brdjanovic et al., 1997) and
through its indirect determination by high-performance
liquid chromatography (HPLC) as glucose after an acid
hydrolysis and extraction (Smolders et al., 1994a;
Lanham et al., 2012). Unfortunately, a direct method is
not yet available. For practical reasons and after several
22
improvements throughout the years, the HPLC method
after acid hydrolysis and extraction is one of the most
frequently applied procedures. The HPLC method, after
acid hydrolysis and extraction for the determination of
glycogen, consists of the digestion of an activated sludge
sample diluted with 6 M HCl, leading to a final HCl
concentration of 0.6 M HCl, and digested at 100 °C for 5
h. After digestion, the sample is allowed to cool down to
room temperature under quiescent conditions and the
supernatant is filtered through 0.2 or 0.45 μm pore size
filters. The filtered supernatant is poured into a vial and
glycogen can be quantified by HPLC as glucose
(Smolders et al., 1994a). The latter is because glycogen
(C6H10O5) shares the same carbon content as glucose
(C6H12O6) (on a carbon mole basis). However, the
determination of glycogen as glucose content extracted
from the biomass is not entirely accurate as the nonglycogen glucose-containing content of the biomass
(cells) will also make a part of the extracted material, and
the glycogen of other glycogen-containing populations
beside EBPR (e.g. GAOs) will do the same too. Recently
Lanham et al. (2012) improved the glycogen extraction
technique. Freeze-dried samples prepared like those for
PHA determination can be used: activated sludge can be
collected in situ, added to a 15 mL centrifugation tube
containing 4-5 drops of formaldehyde (37 %
concentrated), stored at 0-4 °C for around 2 h, washed
with tap water and freeze-dried. They recommend using
a ratio of 1 mg freeze-dried sludge to 1 mL 0.9 M HCl
solution to improve the acid hydrolysis and extraction of
glycogen for its further determination as glucose. Then,
depending upon the sludge aggregation, the sludge
samples can be digested for 2 h in the case of flocculant
sludge, 5 h for granular sludge and 3 h if the aggregation
state is not known or if it varies. Later on, 5 mg of the
freeze-dried sample can be added to 5 mL of a 0.9 M HCl
solution, digested for 5 h at 100 °C, supernatant filtered
through 0.2 μm pore size filters and measured as glucose
by HPLC. If the latter procedure is applied, then the 5 mg
of the freeze-dried sample should be carefully and
precisely weighed and the results will be reported as a
percentage of MLSS. The previous HPLC determination
technique has proven to be sufficiently accurate and
reliable in lab-enriched cultures where EBPR
populations are dominant. However, as previously
discussed, their determination in sludge samples from
full-scale systems may not be accurate enough.
Tentatively, Periodic Acid-Schiff (PAS) stain can be
used to get a rough microscopic qualitative estimation of
glycogen and other carbohydrate granules present in the
cells (Mesquita et al., 2013).
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.2.2.6 Parameters of interest
To determine and assess the metabolic activities of
PAOs, different stoichiometric ratios and kinetic rates for
the anaerobic, anoxic and aerobic stages can be estimated
based on the data collected from the execution of the
batch activity tests. Table 2.2.1 shows a description of the
expected parameters of interest.
2.2.3 EBPR batch activity tests: preparation
This section describes not only the different steps but also
the apparatus characteristics and materials needed for the
execution of the batch activity tests.
2.2.3.1 Apparatus
1. An (airtight) batch bioreactor or fermenter equipped
with a mixing system and adequate sampling ports (as
described in Section 2.2.2.1).
2. A nitrogen gas supply (recommended).
3. An oxygen supply (compressed air or pure oxygen
sources).
4. A pH electrode (if not included/incorporated in the
batch bioreactor setup).
5. A 2-way pH controller via HCl and NaOH addition
(alternatively a one-way control - generally for HCl
addition - or manual pH control can be applied
through the manual addition of HCl and NaOH).
6. A thermometer (recommended temperature working
range of 0 to 40 °C).
7. A temperature control system (if not included in the
batch bioreactor setup).
8. A DO meter with an electrode (if not
included/incorporated in the batch bioreactor setup).
9. An automatic 2-way dissolved oxygen controller via
nitrogen and oxygen gas supplies (if not included in
the batch bioreactor setup and if tests must be
performed at a defined dissolved oxygen
concentration).
10. Confirm that all electrodes and meters (pH,
temperature and DO) are calibrated less than 24 h
before execution of the batch activity tests in
accordance with the guidelines and recommendations
from the manufacturers and/or suppliers.
11. A centrifuge with a working volume capacity of at
least 250 mL to carry out the sludge washing
procedure (if required).
12. A stop watch.
ACTIVATED SLUDGE ACTIVITY TESTS
23
Table 2.2.1 Stoichiometric and kinetic parameters of interest for activated sludge samples performing EBPR.
Parameter
ANAEROBIC PARAMETERS
Stoichiometric
Anaerobic orthophosphate release to VFA uptake ratio
Anaerobic glycogen utilization to VFA uptake ratio
Anaerobic PHA production to VFA uptake ratio
Anaerobic PHB formation to VFA uptake ratio
Anaerobic PHV formation to VFA uptake ratio
Anaerobic PH2MV formation to VFA uptake ratio
Anaerobic PHV formation to PHB formation ratio
Kinetic
Maximum specific anaerobic VFA uptake rate
Maximum specific anaerobic PO4 release rate
Maximum specific anaerobic PHA production rate
Anaerobic PO4 release maintenance rate
Anaerobic ATP maintenance coefficient
Anaerobic secondary PO4 release rate
ANOXIC PARAMETERS
Stoichiometric
Anoxic PHA degradation to NOX consumption ratio
Anoxic glycogen formation to NOX consumption ratio
Anoxic poly-P formation to NOX consumption ratio
Anoxic biomass growth to NOX consumption ratio
Anoxic glycogen formation to PHA consumption ratio
Anoxic poly-P formation to PHA consumption ratio
Anoxic biomass growth to PHA consumption ratio
Kinetic
Maximum specific anoxic PHA degradation rate
Maximum specific anoxic glycogen formation rate
Maximum specific anoxic poly-P formation rate
Maximum specific anoxic biomass growth rate
Anoxic ATP maintenance coefficient
Anoxic endogenous respiration rate
AEROBIC PARAMETERS
Stoichiometric
Aerobic PHA degradation to O2 consumption ratio
Aerobic Glycogen formation to O2 consumption ratio
Aerobic Poly-P formation to O2 consumption ratio
Aerobic PAO biomass growth to O2 consumption ratio
Aerobic glycogen formation to PHA consumption ratio
Aerobic Poly-P formation to PHA consumption ratio
Aerobic biomass growth to PHA consumption ratio
Kinetic
Maximum specific aerobic PHA degradation rate
Maximum specific aerobic glycogen formation rate
Maximum specific aerobic poly-P formation rate
Maximum specific aerobic biomass growth rate
Aerobic ATP maintenance coefficient
Aerobic endogenous respiration rate of a culture
Symbol
Typical unit on a mole basis
Typical unit on a mg or g basis
YVFA_PO4,An
YGly/ VFA,An
YVFA_PHA,An
YVFA_PHB,An
YVFA_PHV,An
YVFA_PH2MV,An
YPHV/PHB,An
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
mg P mg VFA-1
mg C mg VFA-1
mg C mg VFA-1
mg C mg VFA-1
mg C mg VFA-1
mg C mg VFA-1
mg C mg C-1
qVFA,An
qPP_PO4,An
qVFA_PHA,An
mPP_PO4,An
mATP,An
mPP_PO4,Sec,An
C-mol C-mol-1 h-1
P-mol C-mol-1 h-1
C-mol C-mol-1 h-1
P-mol C-mol-1 h-1
mol ATP C-mol-1 h-1
P-mol C-mol-1 h-1
mg VFA mg active biomass-1 h-1
mg P mg active biomass-1 h-1
mg PHA mg active biomass-1 h-1
mg P mg active biomass-1 h-1
mg ATP mg active biomass-1 h-1
mg P mg active biomass-1 h-1
YNOx_PHA,Ax
YNOx_Gly,Ax
YNOx_PP,Ax
YNOx,Bio,Ax
YPHA_Gly,Ax
YPHA_PP,Ax
YPHA_Bio,Ax
C-mol N-mol-1
C-mol N-mol-1
P-mol N-mol-1
C-mol N-mol-1
C-mol C-mol-1
P-mol C-mol-1
C-mol C-mol-1
mg C mg NOX-1
mg C mg NOX-1
mg P mg NOX-1
mg C mg NOx -1
mg C mg C -1
mg P mg C -1
mg C mg C -1
qPHA,Ax
qPHA_Gly,Ax
qPO4_PP,Ax
qBio, Ax
mATP,Ax
mNOx
C-mol C-mol-1 h-1
C-mol C-mol-1 h-1
P-mol C-mol-1 h-1
C-mol C-mol-1 h-1
mol ATP C-mol-1 h-1
N-mol C-mol-1 h-1
mg PHA mg active biomass-1 h-1
mg Gly mg active biomass-1 h-1
mg PP mg active biomass-1 h-1
mg active biomass mg active biomass-1 h-1
mg ATP mg active biomass-1 h-1
mg NOx mg active biomass-1 h-1
YPHA
YGly
YPP
YPAO
YPHA_Gly,Ox
YPHA_PP,Ox
YPHA_Bio,Ox
C-mol mol O2-1
C-mol mol O2-1
P-mol mol O2-1
C-mol mol O2-1
C-mol C-mol-1
P-mol C-mol-1
C-mol C-mol-1
mg C mg O2-1
mg C mg O2-1
mg P mg O2-1
mg C mg O2-1
mg C mg C -1
mg P mg C -1
mg C mg C -1
qPHA,Ox
qPHA_Gly,Ox
qPO4_PP,Ox
qBio, Ox
mATP,Ox
mO2
C-mol C-mol-1 h-1
C-mol C-mol-1 h-1
P-mol C-mol-1 h-1
C-mol C-mol-1 h-1
mol ATP C-mol-1 h-1
mol O2 C-mol-1 h-1
mg PHA mg active biomass-1 h-1
mg Gly mg active biomass-1 h-1
mg PP mg active biomass-1 h-1
mg active biomass mg active biomass-1 h-1
mg ATP mg active biomass-1 h-1
mg O2 mg active biomass-1 h-1
24
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.2.3.2 Materials
1. Two graduated cylinders of 1 or 2 L (depending upon
the sludge volumes used) to hold the activated sludge
and wash the sludge if required.
2. At least 2 plastic syringes (preferably of 20 mL or at
least of 10 mL volume) for the collection and
determination of soluble compounds (after filtration).
3. At least 3 plastic syringes (preferably of 20 mL) for
the collection of solids, particulate or intracellular
compounds (without filtration).
4. 0.45 μm pore size filters. Preferably not of celluloseacetate because they may release traces of cellulose
or acetate into the collected water samples. Consider
having at least twice as many filters as the number of
samples that need to be filtered for the determination
of soluble compounds.
5. 10 or 20 mL transparent plastic cups to collect the
samples for the determination of soluble compounds
(e.g.
soluble
COD,
acetate,
propionate,
orthophosphate, nitrate, nitrite).
6. 10 or 20 mL transparent plastic cups to collect the
samples for the determination of mixed liquor
suspended solids and volatile suspended solids
(MLSS and MLVSS, respectively). Consider the
collection of these samples by triplicate due to the
variability of the analytical technique.
7. 15 mL plastic tubes for centrifugation for the
determination of PHAs and/or glycogen.
8. A plastic box or dry ice box filled with ice with the
required volume to temporarily store (for up to 1-2 h
after the conclusion of the batch activity test) the
plastic cups and plastic tubes for centrifugation after
the collection of the samples.
9. Plastic gloves and safety glasses.
10. Pasteur or plastic pipettes for HCl and/or NaOH
addition (when pH control is carried out manually).
11. Metallic lab clips or clamps to close the tubing used
as a sampling port when samples are not collected
from the bioreactor/fermenter.
2.2.3.3 Media preparation
• Real wastewater
If real wastewater will be used to carry out the batch
activity test, the sample needs to be collected at the
influent of the corresponding wastewater treatment
plant and the batch activity test performed as soon as
possible after collection. Depending on the nature of
the test, the researcher should decide whether to take
a sample of raw sewage or settled sewage (if the plant
employs primary settling). If due to location,
transportation issues or other logistics, tests cannot be
performed in less than 1 or 2 h immediately after
collection, then one should keep the wastewater
sample cold until the test is conducted (e.g. by
placing the bucket or jerry can in a fridge at 4 °C).
Nevertheless, prior to the execution of the test, the
temperature of the wastewater needs to be adjusted to
the target temperature at which the batch activity test
will be executed (preferably reached in less than 1 h).
A water bath or a temperature-controlled room can be
used for this purpose, as described in Section 2.2.2.1.
• Synthetic influent media or substrate
If tests can be or are desired to be performed with
synthetic wastewater, depending on the type of tests
(anaerobic, anoxic or aerobic), the synthetic influent
media can contain a mixture of carbon and
orthophosphate sources plus necessary (macro and
micro) nutrients. Generally, they can be mixed all
together in the same media (for anaerobic-(anoxic)aerobic tests); split in two solutions (i) C source and
(ii) P source (plus nutrient solution); or prepared
separately if they need to be added in different phases
or time. The usual compositions and concentrations
are:
a. Carbon source solution: This is usually composed
of a RBCOD source, preferably volatile fatty
acids such as acetate or propionate, depending on
the nature or goal of the test and the
corresponding research questions. Sometimes,
more complex substrates are used, containing a
mixture of RBCOD and slowly biodegradable
COD (SBCOD); however, these are not applied
in the tests described in this chapter, and thus are
omitted. For anaerobic batch activity tests, the
COD concentration in the feed needs to be set to
a level that ensures that all the COD is consumed
within the anaerobic stage. For batch activity tests
performed with activated sludge from a full-scale
plant, usually COD concentrations not higher
than 100 mg L-1 are recommended. For lab-scale
activated
sludge
samples,
the
COD
concentrations can be as high as the influent COD
concentration of the lab-scale system (and even
sometimes 2 to 3 times higher) as long as the
RBCOD fed is fully removed in the anaerobic
stage and is not toxic or inhibitory to PAOs.
b. Orthophosphate
source
solution:
The
orthophosphate concentrations can be adjusted as
desired depending on the purpose of the
experiment. For single anaerobic batch test
experiments only, orthophosphate concentrations
ACTIVATED SLUDGE ACTIVITY TESTS
can be as low as 2-3 mg PO4-P L-1 or even be
excluded, whereas concentrations as high as 75
mg PO4-P L-1 for full-scale samples or more than
120 mg PO4-P L-1 for lab-enriched cultures
(Wentzel et al., 1987) can be added to assess the
maximum P-uptake capacity of sludge under
anoxic or aerobic conditions (when an anaerobic
phase precedes the anoxic or aerobic test).
c. The nutrient solution: This should contain all the
required macro (ammonium, magnesium,
sulphate, calcium, potassium) and micronutrients
(iron, boron, copper, manganese, molybdate,
zinc, iodine, cobalt) to ensure that cells are not
limited by their absence and avoid obtaining
wrong results and, in extreme cases, the failure of
the test. Thus, despite the fact that their
concentrations may seem very low, it is necessary
to make sure that all of the constituents are added
to the solution in the required amounts. The
following composition (amounts per litre of
nutrient solution) is recommended (based on
Smolders et al., 1994a): 107 mg NH4Cl, 90 mg
MgSO4·7H2O, 14 mg CaCl2·2H2O, 36 mg KCl, 1
mg yeast extract and 0.3 mL of a trace element
solution (that includes per litre 10 g EDTA, 1.5 g
FeCl3·6H2O, 0.15 g H3BO3, 0.03 g CuSO4·5H2O,
0.12 g MnCl2·4H2O, 0.06 g Na2MoO4·2H2O, 0.12
g ZnSO4·7H2O, 0.18 g KI and 0.15 g CoCl·6H2O).
Similar nutrient solutions can be used as long as
they contain all the previously reported required
nutrients.
25
yeast extract and 0.3 mL of a trace element solution
(that includes per litre 10 g EDTA, 1.5 g FeCl3·6H2O,
0.15 g H3BO3, 0.03 g CuSO4·5H2O, 0.12 g
MnCl2·4H2O, 0.06 g Na2MoO4·2H2O, 0.12 g
ZnSO4·7H2O, 0.18 g KI and 0.15 g CoCl·6H2O). The
washing process can be repeated twice or three times.
Afterwards, the following preparation steps of the
batch activity tests can be performed. In special cases
when sludge from a full-scale plant is used, plant
effluent may be used for washing purposes (given
that its composition allows for this).
• Formaldehyde solution
To prepare and preserve the samples for the analytical
determination of PHAs and glycogen, a commercial
formaldehyde solution (37 % concentration) is
needed.
• ATU (Allyl-N-thiourea) solution
To inhibit nitrification, an ATU solution can be
prepared to reach an initial concentration of around
20 mg L-1 (after addition to the sludge). The ATU
solution must be added before the sludge is exposed
to any aerobic conditions (including the sludge
sample preparation or acclimatization).
• Acid and base solutions
• Nitrate or nitrite solution
When the batch tests comprise an anoxic stage,
nitrate or nitrite solutions (as required) can be used to
create the anoxic conditions in the batch activity tests
(Section 2.2.2.4). Nitrate and nitrite salts can be used
for this purpose (e.g. KNO3 or NaNO2, respectively).
Nevertheless, their addition must be carefully
monitored to ensure their presence and availability
without creating any inhibitory or even toxic effect
on the biomass. Thus, it is recommended to keep their
concentrations below 20 mg NO3-N L-1 and 10 mg
NO2-N L-1 (at pH 7.0).
• Washing media
If the sludge sample must be 'washed' to remove an
undesirable compound (which may be even
inhibitory or toxic), it is necessary to prepare a
nutrient solution to wash the sludge that contains per
litre (Smolders et al., 1994a): 107 mg NH4Cl, 90 mg
MgSO4·7H2O, 14 mg CaCl2·2H2O, 36 mg KCl, 1 mg
•
•
These should be 100-250 mL of 0.2 M HCl and 100250 mL 0.2 M NaOH solutions for automatic or
manual pH control, and 10-50 mL of 1 M HCl and
10-50 mL 1 M NaOH solutions for initial pH
adjustment if the desired operational pH is very
different from the pKa value of the buffering agent.
The working and stock solutions required to carry out
the determination of the analytical parameters of
interest must be also prepared in accordance with
Standard Methods (APHA et al., 2012) and the
corresponding protocols.
It is recommended to take a sample of the media prior
to execution of the experiment to confirm/check the
initial (desired) concentration of the parameter(s) of
interest (e.g. COD, orthophosphate, etc.).
2.2.3.4 Material preparation
•
The number of samples (and their volume) to be
collected should be defined in accordance with the
type of analysis, the volumes required for the
analytical determination and the number of replicates
of the dissolved parameters of interest:
26
•
•
•
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
a. E.g. a 3 mL sample needs to be collected if this is
the minimum volume required to determine the
acetate concentration in that sample by duplicate.
b. When two or more parameters will be determined
using the same collected sample, then the
required volumes must be summed up: e.g. if 3
mL is needed for Ac, 5 mL for ammonia, and 6
mL for orthophosphate determination, then at
least a 14 mL sample must be collected.
In a well-mixed system, liquid samples are only taken
once (although the analysis may be conducted in
replicate) because the sampling itself does not usually
affect the quality of the sample (unless the samples
are not filtered by omission).
Also, the number of samples (and their volume) to be
collected should be defined in accordance with the
type of analysis, the volumes or masses required for
the analytical determination of the particular
parameters of interest and the number of replicates
per analysis (e.g. 3 samples of 10 mL each for
MLSS/MLVSS determination by triplicate, 20 mg
TSS per every PHA analysis and 5 mg TSS for each
glycogen analysis). In the case of samples focused on
solids concentrations, there is deviation in the solids
concentrations from sample to sample due to the
nature of sampling procedures. Therefore, the
sampling and analysis preferably need to be
conducted in triplicate.
Frequency of sample collection:
a. If the maximum specific (initial) kinetic rates are
to be determined (e.g. maximum specific acetate
uptake rate or maximum orthophosphate uptake
rate) then a higher number of samples must be
taken at the beginning of the corresponding stage
or phase (anaerobic, anoxic or aerobic). In
particular, samples may need to be collected
every 5 min during the first 30-40 min of duration
of the batch activity test. The 5 min period is the
minimum practical period needed for taking and
handling a single (set of) sample(s).
b. If only the stoichiometric ratios need to be
assessed and not the kinetics (e.g. the anaerobic
P-released/Ac-uptake ratio or anaerobic PHB
formation/Ac uptake ratio), then the samples can
be collected only at the beginning and end of each
stage or phase to determine the total conversions
of interest.
To increase the data reliability and know the initial
conditions of the sludge, it is strongly recommended
to collect a series of sludge samples before any media
is added. In particular for PHAs and glycogen, which
require a long sampling time considering the fast
•
•
•
•
conversion rates of the biomass, similar to MLSS and
MLVSS analyses, samples are often collected in
triplicate together with MLSS and MLVSS samples.
Carefully define the maximum and minimum
working volumes of the bioreactor:
a. The estimation of the minimum final volume at
the end of the test (after the collection of all the
samples) will avoid problems with sampling (e.g.
when the height of the final volume is lower than
the height of the sampling port/tubing) and
controlling the operational conditions (e.g.
uncontrolled acid and base addition will occur if
the tip of the pH electrode is not submerged
leading to a pH shock, without noticing any
problems in the reading of the pH meter). It can
also prevent inadequate or insufficient aeration
and mixing (extremely low volumes can lead to
high oxygen intrusion, dead volumes, and
biomass losses if splashed on the bioreactor
walls).
b. To estimate the minimum initial volume at the
beginning of the test based on the minimum final
volume and the volume required for sampling
(taking into account the initial activated sludge
volume and the addition of media and other
solutions) and to verify that this volume does not
exceed the maximum working volume of the
bioreactor which will allow to avoid spillages and
flooding. The potential increase in volume due to
gas sparging (e.g. nitrogen or compressed air)
must also be considered to define the maximum
working volume.
Once the number and frequency of the sample
collections have been defined, then label all the
plastic cups. Preferably, define a nomenclature and/or
abbreviation that will allow you to easily identify and
recognize the batch test, the sampling time and the
parameter(s) of interest to be determined with that
sample. Labelling both the plastic cups and the cover
will help to easily identify the sample.
A simple working sheet created in a spreadsheet can
be rather useful to execute and keep track of the
sampling collection and batch test execution.
Furthermore, it can be used to keep a database of the
different batch tests carried out. Table 2.2.10 contains
an example of a working sheet.
Organize all the required material within a relatively
close radius of action around the batch setup so any
delay in handling and preparing the samples can be
avoided. Otherwise, it will be difficult to respect the
initial 5 min frequency of sampling.
ACTIVATED SLUDGE ACTIVITY TESTS
•
27
If samples for PHA and glycogen determination are
to be collected, carefully add 4-5 drops of the 37 %
formaldehyde solution using a plastic Pasteur pipette.
Add the formaldehyde solution inside a fume hood or
at least in a well ventilated place. After addition,
close the tubes immediately and keep them closed
until the samples are added. Always wear plastic
gloves and treat the used materials contaminated with
formaldehyde as chemical residue in accordance with
your local lab regulations.
•
•
Calibrate all the meters (pH, DO and thermometer)
less than 24 h prior to the execution of the tests and
store the electrode/sensors in appropriate solutions
until the execution of the tests, following the
particular recommendations of the corresponding
manufacturer or supplier and confirm that the
readings are reliable.
One should be aware that each sample taken needs
immediate attention, handling and proper storage,
before the next sample is taken (Table 2.2.2).
Table 2.2.2 Suggestions for the storage and preservation of samples as a function of the analytical determination of the parameter of interest.
EBPR-related parameter of
interest
Material of sample
container
Method of preservation
Maximum recommended time between sampling,
preservation procedure and analysis
Total BOD
Plastic or glass
Soluble/dissolved BOD
Plastic or glass
24 h for samples stored at 1-5 °C;
up to 1 month for frozen samples.
24 h for samples stored at 1-5 °C;
up to 1 month for frozen samples.
Total COD
Plastic or glass
Soluble/dissolved COD
Plastic or glass
Cool to 1-5 °C;
or, freeze to -20 °C and store in the dark.
Filter immediately after collection through 0.45 μm
pore size filters and cool to 1-5 °C;
or, freeze to -20 °C and store in the dark.
Cool to 1-5 °C;
or, add concentrated H2SO4 to lower pH to 1-2, freeze
to -20 °C and store in the dark .
Filter immediately after collection through 0.45 μm
pore size filters, cool to 1-5 °C.
VFA
Plastic or glass
Total P
NH4
Plastic or glass acid
washed (0.1 M HCl)
Plastic or glass acid
washed (0.1 M HCl)
Plastic or glass
NO2
Plastic or glass
NO3
Plastic or glass
PHAs
Plastic
Glycogen
Plastic or glass
MLSS
MLVSS
Plastic or glass
Plastic or glass
PO4
Filter immediately after collection through 0.45 μm
pore size filters, cool to 1-5 °C;
or, add concentrated H2SO4 to lower pH to 1-2, freeze
to -20 °C and store in the dark.
Add concentrated H2SO4 to lower pH to 1-2 and freeze
to -20 °C.
Filter immediately after collection through 0.45 μm
pore size filters and cool to 1-5 °C.
Filter immediately after collection through 0.45 μm
pore size filters, cool to 1-5 °C;
or, add concentrated H2SO4 to lower pH to 1-2, freeze
to -20 °C and store in the dark.
Filter immediately after collection through 0.45 μm
pore size filters, cool to 1-5 °C.
Filter immediately after collection through 0.45 μm
pore size filters, cool to 1-5 °C;
or, add concentrated HCl to lower pH to 1-2, freeze to
-20 °C and store in the dark.
After corresponding sampling and preparation
procedure (see Section 2.2.2.5) store at -20 °C or -80
°C; freeze-dried samples can also be stored at -20 or 80 °C.
After corresponding sampling and preparation
procedure (see Section 2.2.2.5) store at -20 °C;
after digestion also store at -20 °C.
Cool to 1- 5 °C.
Cool to 1-5 °C.
24 h for samples stored at 1-5 °C;
up to 6 months for acidified samples frozen and
stored in the dark.
24 h for samples stored at 1-5 °C;
up to 6 months for acidified samples frozen and
stored in the dark.
24 h for samples stored at 1-5 °C;
up to 6 months for acidified samples frozen and
stored in the dark.
6 months.
24 h.
24 h for samples stored at 1-5 °C;
up to 21 days for acidified samples frozen and stored
in the dark.
24 h.
24 h for samples stored at 1-5 °C;
7 days for acidified samples frozen and stored in the
dark.
Up to 6 months.
Up to 6 months.
2 days.
24 h.
28
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.2.3.5 Activated sludge preparation
These procedures consider that batch activity tests can be
performed as soon as possible after the collection of
samples from full- or lab-scale systems or, in the worstcase scenario, within 24 h after collection. The execution
of batch tests 24 h after the collection of activated sludge
samples is not recommended due to potential changes
that EBPR culture can experience during handling
(unless the exposure time after collection is of particular
interest for the execution of tests). Bearing in mind the
previous comments, the following three procedures are
recommended to prepare the activated sludge samples for
the execution of batch activity tests:
• If batch activity tests can be executed in less than 1 h
after collection of the sludge sample and if the sludge
sample does not need to be washed:
a. Adjust the temperature of the batch bioreactor
where the tests will take place to the target
temperature of the study.
b. Collect the sludge:
i. At the end of the aerobic tank or stage to carry
out anaerobic batch tests.
ii. At the end of the anaerobic stage or tank to
perform anoxic or aerobic batch tests.
c. Transfer the sludge sample to the bioreactor or
fermenter where the batch activity tests will take
place.
d. Add the ATU solution (if applicable for the
objective of the test) to a final concentration of 20
mg L-1 (see Section 2.2.3.3).
e. Start a gentle mixing (50 - 100 rpm) and follow
the temperature of the sludge sample by placing
an external thermometer inside the bioreactor (if
the setup does not have a built-in thermometer).
f. Keep mixing until the sludge has reached the
target temperature of the study.
g. Keep the same redox conditions prevailing during
collection until the batch activity tests are ready
to start:
i. Avoid the aeration of samples collected under
anaerobic conditions. If available, use an
airtight bioreactor and sparge nitrogen gas to
avoid/reduce oxygen intrusion.
ii. Avoid the aeration of samples collected under
anoxic conditions and preferably add a nitrate
solution to a final concentration of around
10 mg NO3-N L-1 to preserve the anoxic
environment while mixing.
iii. Aerate the sludge samples collected in the
aerobic tank or under aerobic conditions,
•
keeping a dissolved oxygen concentration
higher than 2 mg L-1.
h. Sludge for the execution of anaerobic batch tests:
if nitrate is detected in the aerobic samples
collected to perform these tests (see Section
2.2.2.3), the aeration must stop and the sludge
sample should be gently mixed until nitrate is no
longer observed. As soon as nitrate is no longer
observed, the sludge can be immediately used to
execute the anaerobic batch activity tests. Nitrate
or nitrite detection strips (Sigma-Aldrich) can be
used for a quick estimation of the presence of
these compounds.
i. Sludge to perform anoxic and aerobic batch
activity tests can then be used to carry out the tests
within 1 h of collection.
If batch activity tests can be conducted in less than 1
h after collection of the sludge sample but the
activated sludge sample needs to be washed:
a. Adjust the temperature of the batch bioreactor
where the tests will take place to the target
temperature of the study. Collect the sludge at the
end of the aerobic tank or stage.
b. Wash the sludge in a mineral solution (see
Section 2.2.3.3) as follows:
i. Separate the biomass by settling or mild
centrifugation (2,000-3,000 rpm for 5 min)
and carefully remove the supernatant volume
while avoiding losing the biomass.
ii. Replace the supernatant volume with the
same volume of nutrient solution and mix
gently for 5 min.
iii. Repeat the previous washing procedure at
least one more time.
iv. After the last washing cycle, separate the
biomass by settling or mild centrifugation
(2,000-3,000 rpm for 5 min) and re-suspend
the sludge in the same volume of mineral
solution previously added.
v. The washed sample should have the same
MLSS concentration as in the lab- or fullscale system where it was taken from. Thus,
define and adjust the volume of the mineral
solution added to re-suspend the washed
sludge in order to reach the same MLSS
concentration like in the original source.
c. Transfer the washed sludge sample to the
bioreactor or fermenter where the batch activity
tests will take place.
d. Add the ATU solution (if applicable for the
objective of the tests) to a final concentration of
20 mg L-1 (see Section 2.2.3.3).
ACTIVATED SLUDGE ACTIVITY TESTS
•
e. Start a gentle mixing (50-100 rpm) and follow the
temperature of the sludge sample by placing an
external thermometer inside the bioreactor (if the
setup does not have a built-in thermometer).
f. Keep mixing until the sludge is exposed to the
target temperature of the study for at least 30 min.
g. Start to aerate the sludge sample, keeping a
dissolved oxygen concentration not lower than 2
mg L-1 while providing a gentle mixing until the
batch activity test starts.
h. Due to the exposure of the sludge to aerated and
non-aerated conditions, it is only recommended
to use this sludge to execute batch activity tests
that start with an anaerobic phase (anaerobic(anoxic)-aerobic). The execution of batch activity
tests that start with an anoxic or aerobic phase is
not recommended because the washing steps may
decrease the intracellular PHA contents and polyP contents due to handling during the washing
steps.
If due to location and distance issues, the tests cannot
be performed within less than 1 or 2 h after collection
(but within 24 h):
a. Adjust the temperature of the batch bioreactor to
the target temperature of the study.
b. Keep the sludge sample cold until the test is
executed (e.g. by placing the bucket or jerry can
in a fridge at 4 °C), avoiding aerating the sludge
sample.
c. Prior to the execution of the test, take the sludge
sample out of the fridge, cool box or cold room.
d. Mix the content gently in order to obtain a
homogenous and representative sample with a
similar MLSS concentration as in the original labor scale system where it was collected from.
e. If the sample needs to be washed, wash the sludge
in a mineral solution (see Section 2.2.3.3) as
follows (otherwise, skip the washing step):
i. Separate the biomass by settling or mild
centrifugation (2,000-3,000 rpm for 5 min)
and carefully remove the supernatant volume
while avoiding losing the biomass.
ii. Replace the supernatant volume with the
same volume of nutrient solution and mix
gently for 5 min.
iii. Repeat the previous washing procedure at
least one more time.
iv. After the last washing cycle, separate the
biomass by settling or mild centrifugation
(2,000-3,000 rpm for 5 min) and re-suspend
the sludge in the same volume of nutrient
solution previously added.
29
v.
f.
g.
h.
i.
j.
k.
l.
Since the washed sample should have the
same MLSS concentration as in the lab- or
full-scale system where it was taken from,
define and adjust the volume of the mineral
solution to obtain the same MLSS
concentration as in the original source.
Transfer the washed sludge to the bioreactor or
fermenter where the batch activity test will take
place.
Add the ATU solution to a final concentration of
20 mg L-1 (see Section 2.2.3.3).
Start to aerate the sludge sample, keeping the
dissolved oxygen concentration higher than 2 mg
L-1 while mixing gently.
Follow the temperature of the sludge sample by
placing an external thermometer inside the
bioreactor (if the setup does not have a built-in
thermometer).
Keep aerating and mixing for at least 1 h
(maximum 2 h) but ensure that the sludge is
exposed to the target temperature of the study for
at least 30 min.
If nitrate is detected after the procedure to adjust
the temperature, (see Section 2.2.2.3 on activated
sludge sample preparation), the aeration must
stop. Mix gently until nitrate is no longer
observed. Afterwards, the sludge can be
immediately used to start and execute an
anaerobic phase test.
If the sample was 'washed', it is only
recommended to use this sludge to execute batch
activity tests that start with an anaerobic phase
(e.g. anaerobic-(anoxic)-aerobic), because the
washing steps may decrease the intracellular PHA
contents and poly-P contents due to handling
during the washing steps. Thus, an anaerobic
stage is always needed to replenish the PHA
contents of the biomass.
2.2.4 Batch activity tests: execution
Once the experimental setup, materials, solutions, and
activated sludge are ready, the corresponding batch
activity test can be conducted. To facilitate the execution
and for data track record and archiving purposes, an
experimental implementation plan should be prepared in
advance. Table 2.2.10 presents a template for an
experimental implementation plan that can be used with
necessary modifications for the execution of each of the
batch activity tests described in the following sections.
Due to the particular metabolism of EBPR cultures,
EBPR batch activity tests can range from anaerobic to
30
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
anoxic and aerobic tests, including different
combinations among them, depending on the purpose or
goal of the test.
Test code no.
EBPR.ANA.1
Redox conditions
Anaerobic
EBPR.ANA.2
Anaerobic
EBPR.ANA.3
Anaerobic
EBPR.ANOX.1
Anoxic
EBPR.ANOX.2
Combined anaerobic-anoxic
EBPR.AER.1
Aerobic
EBPR.AER.2
Combined anaerobic-aerobic
EBPR.AER.3
Combined anaerobic-anoxic-aerobic in series
EBPR.AER.4
Combined anaerobic-anoxic-aerobic in parallel
2.2.4.1 Anaerobic EBPR batch activity tests
The length of an anaerobic batch test can last from 1 h to
more than 8 h. Biomass is sensitive to pH and
temperature, so the tests should be conducted at the
temperature and pH of interest and fluctuations should be
avoided. Tests should be executed under the absence of
any electron acceptor (molecular oxygen, nitrate or
nitrite) (e.g. truly anaerobic conditions). To avoid or
minimize oxygen intrusion, it is recommended to use
airtight reactors/fermenters and, if available, sparge N2
gas continuously throughout the execution of the test. To
remove nitrate present in the sample, ATU must be added
during the sample preparation (before the sample is
aerated) and the sludge can be gently mixed for a few
minutes to remove any residual nitrate. Depending upon
the purpose of the anaerobic batch activity test, the
availability and presence of electron donors can vary.
The following anaerobic tests are widely performed:
1. Test EBPR.ANA.1 Performed under the absence of
external carbon source to assess the endogenous
anaerobic maintenance activity of EBPR cultures.
2. Test EBPR.ANA.2 Carried out after the addition of a
defined concentration of carbon (which should be
Thus, the following EBPR batch activity tests are
presented in this chapter:
Short description and purpose
Executed under the absence of an external carbon source to assess the
endogenous anaerobic maintenance activity of EBPR cultures.
Performed after the addition of a defined concentration of carbon to determine
the maximum anaerobic activity of EBPR cultures.
Carried out after the addition of a carbon source in excess to estimate the
maximum activity of EBPR cultures under non-limiting carbon conditions.
Performed with activated sludge samples collected at the end of the anaerobic
stage/phase.
Anaerobic-anoxic test conducted with sludge collected at the end of an aerobic
stage.
Performed with sludge collected at the end of an anaerobic or anoxic stage to
assess the aerobic EBPR activity.
Anaerobic-aerobic test conducted with sludge collected at the end of an
aerobic stage.
Anaerobic-anoxic-aerobic test carried out with sludge collected at the end of
an aerobic stage to assess the sequential anaerobic, anoxic and aerobic EBPR
activities.
After the conduction of a common anaerobic phase, one anoxic and one
aerobic test are performed in parallel with the same sludge to assess the
anoxic and aerobic EBPR activities for comparison purposes.
fully consumed within the duration of the anaerobic
test): to determine the maximum carbon uptake rate,
maximum P-release rate, half-saturation constant for
carbon
uptake,
and
associated
anaerobic
stoichiometry such as P-released to carbon consumed
ratio.
3. Test EBPR.ANA.3 Executed after the addition of a carbon
source in excess of the concentration that an EBPR
culture could consume within the duration of the test:
to estimate the maximum concentration of
phosphorus that can be released and the maximum
concentration of carbon that the EBPR cultures can
consume under non-limiting carbon conditions.
Since the presence and type of carbon source play a
major role in EBPR processes, sludge collection and
preparation are of major importance. It is preferable to
execute these tests with activated sludge collected at the
end of the aerobic stage (to minimize the presence of an
originally present carbon source) and/or after following a
washing procedure. Thus, the following protocols for the
execution of anaerobic batch tests are suggested
depending upon the presence and availability of an
external carbon source:
ACTIVATED SLUDGE ACTIVITY TESTS
Test EBPR.ANA.1 Anaerobic batch EBPR tests performed under
the absence of an electron donor
a. After sludge has been collected, prepared and
transferred to the batch bioreactor (see Section
2.2.3.5), keep the sample aerated for at least 30 min
while confirming that the pH and temperature are at
the target value of interest. Otherwise, set up the
corresponding set points (if automatic pH and
temperature controllers are applied) or adjust
manually. Wait until stable conditions are reached.
b. Once stable operating conditions are reached, around
20 min before the start of the test take the first
samples of the water phase and biomass to determine
the initial concentrations of the parameters of
interest: C-source, total P, PO4 and MLSS and
MLVSS
concentrations.
Samples
for
the
determination of PHAs and glycogen can also be
collected to assess the anaerobic stoichiometric
conversions. It is recommended to also take samples
of the media to check and verify the initial
concentrations.
c. For sampling, connect the syringe, open or release the
lab clip or clamp that closes the sampling port, and
pull and push the syringe several times until a
homogenous sample is collected (usually around 5
times are required). Next, when the syringe is full,
close the clip and remove the syringe.
d. Samples for the determination of soluble components
must be immediately filtered (through 0.45 μm pore
size filters). Other samples (e.g. PHAs, glycogen)
need to be prepared in accordance with the
corresponding protocols explained earlier in the
chapter.
e. During the test execution, temporarily store the
samples at 4 °C in the fridge or preferably in a cool
box with ice.
f. 10 min before the start of the test stop the aeration
and close the bioreactor.
g. If available, start to sparge N2 gas and continue
sparging until the end of the batch test. Provide an
adequate outlet to allow the exit of N2 gas and avoid
building up overpressure (which can lead to flooding
and biomass loss). If a continuous N2 gas sparging
throughout the batch test is not feasible, sparge N2 gas
for 10 min (alternatively another suitable gas could
be used to remove the oxygen present and avoid its
intrusion) and afterwards keep the bioreactor under
airtight conditions.
h. Start the execution of the anaerobic test at ‘time zero’.
Keep track of the execution time with the stopwatch.
i. Duration and sampling:
31
i.
If only the anaerobic endogenous maintenance Prelease rate must be determined, the test can last
for 1 h or maximum 2 h with continuous sampling
every 15 min for the determination of PO4
released by biomass (PAOs) under the absence of
external carbon.
ii. If the anaerobic endogenous maintenance
glycogen conversion rate and stoichiometric
conversions are also of interest, the anaerobic
tests must be extended for up to 6 and 8 h
(recommended). Samples for PO4 determination
can be collected every 30 min together with
samples for PHA and glycogen determination.
iii. Conclude the anaerobic test with the collection of
samples for the determination of COD, Total P,
PO4 and MLSS and MLVSS concentrations, as
well as for PHAs and glycogen (if applicable).
j. Ensure that considerable temperature and pH
variations (higher than 1 °C or ± 0.1, for temperature
and pH, respectively) do not take place during the
execution of the batch test, and that DO readings
remain below the detection limits. Make sure that all
the electrodes used are recently calibrated before the
execution of the test.
k. Organize the samples and ensure that all the samples
are complete and properly labelled to avoid mixing of
samples and other trivial mistakes.
l. Until the collected samples are analysed, preserve
and store them as recommended by the corresponding
analytical procedures.
m. Clean up the apparatus and take appropriate measures
to keep and preserve the different sensors, equipment
and materials.
n. Keep (part of) the sludge used in the test for possible
further use (e.g. for microbial identification, see
chapters 7 and 8).
Test EBPR.ANA.2 Anaerobic batch EBPR tests performed under
a defined addition of an electron donor
a. Repeat steps ‘a’ to ‘g’ from Test EBPR.ANA.1.
b. For anaerobic tests performed with the presence of an
external carbon source (electron donor), the tests start
at 'time zero' with the addition of the real or synthetic
wastewater (as a carbon source solution).
c. To execute the tests, the following MLVSS and
RBCOD concentrations are suggested depending on
the origin of the sludge samples:
i. Sludge samples from full-scale activated sludge
systems: add the RBCOD source to reach an
initial RBCOD-to-MLVSS ratio in the bioreactor
of between 0.025 and 0.050 mg COD mg VSS-1.
32
For instance, mix the RBCOD source with the
fresh activated sludge in such a way that the initial
RBCOD concentration in the bioreactor is
between 50 and 100 mg COD L-1 and the initial
MLVSS concentration is around 2,000 mg VSS
L-1. A low RBCOD-to-MLVSS ratio is preferable
to ensure the RBCOD consumption for the
duration of the anaerobic test.
ii. Sludge samples from lab-scale activated sludge
systems: the RBCOD source can be added to
reach an initial RBCOD-to-MLVSS ratio in the
bioreactor of between 0.05 and 0.10 mg COD mg
VSS-1. For example, the initial RBCOD and
MLVSS concentrations after mixing can range
between 100 and 300 mg COD L-1 and 2,000 and
3,000 mg VSS L-1, respectively. Higher
concentrations may also be acceptable as long as
the COD is fully consumed within the length of
the anaerobic stage. However, avoid COD
concentrations higher than 800 mg COD L-1
because this may be inhibitory to biomass
(author's personal observations).
d. After the addition of the wastewater, keep track of the
execution and sampling times with a stopwatch.
e. Duration and sampling:
i. Tests can last between 2 and 4 h.
ii. To determine the anaerobic kinetic parameters,
samples for the determination of soluble COD
and PO4 should be collected every 5 min in the
first 30-40 min of execution of the test. After this
period, the sampling frequency can be reduced to
10 or 15 min during the first 1 h, and later on to
every 15 or 30 min until the test is finished.
iii. If the anaerobic kinetic conversion of PHAs and
glycogen is of interest, samples should be
collected at the same time as the samples for COD
and PO4 determination.
iv. For stochiometric conversions, samples for PHA
and glycogen determination must be taken at the
beginning and end of the test.
v. Conclude the anaerobic test with the collection of
samples for the determination of C-source and/or
COD (as desired depending upon the parameters
of interest), total P, PO4 and MLSS and MLVSS
concentrations, as well as for PHAs and glycogen
(if applicable).
f. Repeat steps ‘j’ to ‘n’ from Test EBPR.ANA.1.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Test EBPR.ANA.3 Anaerobic batch EBPR tests performed after
the addition of an electron donor in excess
a. Repeat steps ‘a’ to ‘g’ from Test EBPR.ANA.1.
b. For anaerobic tests performed with the presence of an
external carbon source (electron donor), the tests start
at 'time zero' with the addition of the real or synthetic
wastewater (as a carbon source solution).
c. To execute the tests, the following MLVSS and
RBCOD concentrations are suggested depending on
the origin of the sludge samples:
i. Sludge samples from full-scale activated sludge
systems: Perform the anaerobic tests with an
initial RBCOD-to-MLVSS ratio in the bioreactor
higher than 0.15 mg COD mg VSS-1 after mixing
the RBCOD source and the sludge. For instance,
for sludge samples with MLVSS concentrations
of around 2,000, add 300 mg COD L-1 of the
RBCOD source of interest. Higher concentrations
can be added but avoid adding more than 800 mg
COD L-1 since this has been shown to be
inhibitory for EBPR cultures (author's personal
observation). If all RBCOD is consumed, more
RBCOD can be added until it is not totally
consumed. Monitoring the PO4 concentration’s
profile during the execution of the test can be used
as an indirect method to assess whether the sludge
has
reached
its
maximum
RBCOD
removal/uptake capacity. This can be applied if
after an additional dose of RBCOD source there
is not a considerable increase in PO4
concentrations (e.g. less than 2-3 mg PO4-P L-1
during 30-60 min, which corresponds to
anaerobic endogenous P release).
ii. Sludge samples from lab-scale activated sludge
systems: similarly to full-scale sludge samples,
apply an initial RBCOD-to-MLVSS ratio in the
bioreactor higher than 0.2 mg COD mg VSS-1
(after mixing the COD source and the sludge).
Lab-scale cultures, particularly from EBPR
systems, have a considerably high RBCOD
removal capacity, which may require repeating
the COD addition more than twice until no further
COD uptake is observed. For instance, for sludge
samples with a MLVSS concentration of between
2,000 and 3,000 mg L-1, add at least 400 mg COD
L-1 of the RBCOD source of interest, but avoid
adding more than 800 mg COD L-1 (due to the
potentially inhibitory effects on EBPR cultures).
If all RBCOD is consumed, more RBCOD can be
added until a residual COD is observed.
Monitoring the PO4 concentrations profile during
ACTIVATED SLUDGE ACTIVITY TESTS
the execution of the test can be used as an indirect
method to assess whether the sludge has reached
its maximum RBCOD removal capacity. This
approach can be applied if after an additional dose
of RBCOD source there is not a considerable
increase in PO4 concentrations (e.g. less than 2-3
mg PO4-P L-1 after 30-60 min).
d. Start to keep track of the execution time with a
stopwatch just after the addition of the
wastewater.
e. Duration and sampling:
i. Tests can last more than 2-4 h for samples
from full-scale systems and even longer for
lab-scale systems depending on the poly-P
and glycogen content of the biomass.
ii. To determine the anaerobic kinetic
parameters, samples for the determination of
C-source and/or COD (depending upon the
parameters of interest) and PO4 must be
collected every 5 min in the first 30 min of
execution of the test. After this period, the
sampling frequency can be reduced to every
10 or 15 min during the first 1 h, and later on
to every 15 or 30 min until the test is finished.
iii. If the anaerobic kinetic conversions of PHAs
and glycogen are of interest, samples can be
collected at the same time as the samples for
C-source and/or COD and PO4 determination.
iv. For stochiometric conversions, samples for
the determination of PHAs and glycogen must
be taken at the beginning and end of the test.
v. Prior to the end of the tests, add an additional
concentration of COD. Wait for 10-15 min
and take the last sample for COD and PO4
determination. If no additional COD uptake
and P release are observed, this will help to
confirm whether the test was indeed
performed
under
non-limiting
COD
conditions.
vi. Conclude the anaerobic test with the
collection of samples for the determination of
C-source and/or COD, total P, PO4 and MLSS
and MLVSS concentrations, as well as for
PHAs and glycogen (if applicable).
f. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
2.2.4.2 Anoxic EBPR batch tests
Anoxic EBPR batch tests are conducted to assess the
simultaneous removal of orthophosphate and nitrate (or
nitrite) by PAOs. They can be performed with activated
sludge samples from full- or lab-scale systems, using real
33
or synthetic wastewater/solutions. It is important that the
EBPR biomass has enough intracellularly stored PHAs
available when exposed to anoxic conditions as a carbon
and energy source for P-uptake, glycogen formation,
biomass growth and maintenance (Figure 2.2.1). In any
case, no external carbon source should be added. Sludge
samples can be collected at the end of the anaerobic stage
or phase, as long as RBCOD has been fully consumed
(and therefore absent from the activated sludge). To
ensure the availability of PHAs, it is strongly advised to
avoid washing the activated sludge or biomass in
between the sludge collection and handling at the end of
the anaerobic stage and the start of the anoxic batch test.
Under exceptional circumstances (e.g. the suspected
presence of toxic compounds), activated sludge samples
may be washed as long as strictly anaerobic conditions
can be created during the washing procedure (a
complicated procedure in practice). Instead, the sludge
can be collected at the end of the aerobic stage and the
anoxic EBPR test can start with a preceding anaerobic
batch test under the defined addition of an electron donor
(similar to Test EBPR.ANA.2, described in Section
2.2.4.1). Thus, the following anoxic activity batch tests
are suggested:
1. Test EBPR.ANOX.1 Single anoxic EBPR test performed
with activated sludge samples collected at the end of
the anaerobic stage/phase.
2. Test EBPR.ANOX.2 Combined anaerobic-anoxic EBPR
batch test. The anoxic EBPR batch test is conducted
after a preceding anaerobic phase with sludge
collected at the end of the aerobic stage.
The description of each test is presented below:
Test EBPR.ANOX.1 Single anoxic EBPR batch tests
a. Collect the activated sludge sample in the full- or
lab-scale plant at the end of the anaerobic stage.
Prepare and transfer it to the batch bioreactor as
described in Section 2.2.3.5. Keep the sample
under anaerobic conditions for at least 30 min
while confirming that the pH, DO and
temperature are at the target values of interest
(otherwise adjust and wait until stable conditions
are reached).
b. Repeat steps 'b' to 'e' from Test EBPR.ANA.1.
c. If available, sparge N2 gas continuously until the
end of the batch test. Provide an adequate outlet
to allow the exit of N2 gas and avoid any
overpressure. If continuous N2 gas sparging
throughout the batch test cannot be applied,
sparge N2 gas for 10 min (alternatively another
34
suitable gas can be used to remove the oxygen
present and avoid its intrusion) and afterwards
keep the bioreactor airtight.
d. Start the execution of the anoxic EBPR test at
'time zero' with the addition of nitrate or nitrite
(depending upon the final electron acceptor of
interest). The same operating conditions like
those applied for the anaerobic tests must be
applied to avoid the intrusion of oxygen (see Test
EBPR.ANA.1).
e. Duration and sampling of the anoxic EBPR batch
test:
i. The anoxic EBPR test can last between 2 and
4 h.
ii. Depending upon the final electron acceptor of
interest:
• Anoxic EBPR tests performed with nitrate (NO3-) as the final
electron acceptor:
For activated sludge samples exposed to the
presence of nitrate, up to 20 mg NO3-N L-1
can be added at the beginning of the anoxic
stage. On the other hand, it is not
recommended to add more than 20 mg NO3N L-1 to activated sludge samples not
regularly
exposed
to
high
nitrate
concentrations. Nevertheless, if all the nitrate
is consumed, an additional 20 mg NO3-N L-1
can be added to extend the length of the
anoxic stage until no further anoxic P-uptake
is observed.
• Anoxic EBPR tests performed with nitrite (NO2-) as the final
electron acceptor:
Usually, activated sludge samples are not
exposed to high nitrite concentrations, unless
they are acclimatized to the presence of this
electron acceptor (in exceptional/particular
cases). Thus, for activated sludge samples
exposed to the presence of nitrite, up to 20 mg
NO2-N L-1 can be added at the beginning of
the anoxic stage. On the other hand, it is not
recommended to add more than 10 mg NO2N L-1 to activated sludge samples not
regularly
exposed
to
high
nitrite
concentrations. In the latter case, if all the
nitrite is consumed, an additional 10 mg NO2N L-1 can be added to ensure the availability
of electron acceptor and the anoxic stage can
be extended until no further anoxic P-uptake
is observed.
iii. Since the analytical determination of nitrate or
nitrite can not be determined as quickly as needed
to monitor their presence during the test, nitrate
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
and/or nitrite detection strips (Sigma-Aldrich)
can be used to rapidly estimate their presence and
to a certain degree of accuracy their
concentration, and to assess whether the anoxic
conditions are still present or if additional nitrate
or nitrite must be dosed.
iv. For the estimation of the anoxic kinetic
parameters, samples for the determination of Csource (or soluble COD depending upon the
analytical parameter of interest) and PO4 must be
collected every 5 min in the first 30 min of
execution of the test. After this period, the
sampling frequency can be reduced to every 10 or
15 min during the first 1 h, and later on to every
15 or 30 min until the test is finished.
v. If the anoxic kinetic conversions of PHAs and
glycogen are of interest, samples can be collected
at the same time as the samples for PO4
determination.
vi. To estimate the anoxic stochiometric
conversions, samples for the analytical
determination of total P, PO4, NO3 (or NO2, and
MLSS and MLVSS concentrations, as well as for
PHAs and glycogen (if applicable), must be
collected both at the start and at the end of the
anoxic phase.
f. Conclude the anoxic test with the collection of the
last samples needed for the estimation of the anoxic
stoichiometric conversions.
g. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
Test EBPR.ANOX.2 Combined anaerobic-anoxic EBPR batch
tests
a. Repeat steps 'a' to 'g' from Test EBPR.ANA.1 and
steps 'b' to 'f' from Test EBPR.ANA.2. Afterwards,
continue with the execution of the anoxic stage.
b. Immediately after the anaerobic stage is completed,
the anoxic EBPR batch test can start immediately
with the addition of nitrate or nitrite (Test
EBPR.ANOX.1, step ‘d’). The same operating
conditions like those applied for anaerobic tests (e.g.
Test EBPR.ANA.1) must be applied to avoid the
intrusion of oxygen.
c. Duration and sampling of the anoxic EBPR batch
test: repeat step 'e' from Test EBPR.ANOX.1.
d. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
2.2.4.3 Aerobic EBPR batch tests
Aerobic EBPR batch tests can be executed to assess the
orthophosphate uptake by PAOs in activated sludge
ACTIVATED SLUDGE ACTIVITY TESTS
systems. They can be performed with activated sludge
samples from full- or lab-scale systems, using real or
synthetic wastewater/solutions. Similar to anoxic BPR
tests, it is important that the EBPR biomass has enough
intracellularly stored PHAs available when exposed to
aerobic conditions as a carbon and energy source for Puptake, glycogen formation, biomass growth and
maintenance (Figure 2.2.1). To ensure the availability of
intracellularly stored PHAs, aerobic batch tests must be
(i) performed with activated sludge samples collected at
the end of the anaerobic or anoxic stage/phase
(depending upon the wastewater treatment plant
configuration), or, (ii) carried out after an anaerobic batch
test and before the aerobic test is conducted. In any case,
no external carbon source should be added during the
aerobic phase. Thus, samples can be collected at the end
of the anaerobic stage or phase, as long as RBCOD is not
present. Alternatively, the aerobic test can be performed
in a sequential mode after an anaerobic-anoxic test (Test
EBPR.ANOX.1) resulting in an anaerobic-anoxicaerobic test or, in parallel to an anoxic test after the
execution of an anaerobic test (Test EBPR.ANA.2)
(Wachtmeister et al., 1997). It is strongly recommended
to avoid washing the activated sludge or biomass in
between the sludge collection and handling from the
anaerobic or anoxic stage and the start of the aerobic
batch test to reduce the potential oxidation of the PHAs.
If the biomass needs to be washed (e.g. due to the
suspected presence of toxic compounds), then an
anaerobic test should always be executed prior to the
aerobic batch test under the defined presence of an
electron donor (Test EBPR.ANA.2). Thus, the following
aerobic EBPR batch tests (with and without preceding
anaerobic or anoxic stages) can be proposed:
1. Test EBPR.AER.1 A single aerobic EBPR test performed
with sludge collected at the end of an anaerobic or
anoxic stage to assess the aerobic EBPR activity on
the intracellular polymers stored in the original
source of sludge.
2. Test EBPR.AER.2 Combined anaerobic-aerobic EBPR
batch tests conducted to ensure a defined
concentration of intracellular PHAs to cover the
aerobic metabolic requirements using sludge
collected at the end of an aerobic stage.
3. Test EBPR.AER.3 Combined anaerobic-anoxic-aerobic
EBPR batch tests in series carried out with sludge
collected at the end of an aerobic stage to assess the
sequential anaerobic, anoxic and aerobic EBPR
activities after securing a defined concentration of
intracellularly stored PHAs.
35
4. Test EBPR.AER.4 Combined anaerobic-anoxic-aerobic
EBPR batch tests in parallel performed with sludge
collected at the end of an aerobic stage to assess the
anoxic and aerobic EBPR activities in parallel after
an anaerobic phase. Since the anaerobic test is
common, the sludge has the same content of
intracellularly stored polymers at the beginning of the
anoxic and aerobic tests, so both anoxic and aerobic
EBPR activities can be compared to each other and in
some cases even conducted in parallel (usually the
privilege of more experienced experimenters).
The description of the different steps involved in the
execution of aerobic EBPR tests is as follows:
Test EBPR.AER.1 Single aerobic EBPR test
a. Collect the sludge in the anaerobic stage or anoxic
stage following the recommendations for sampling
and activated sludge preparation described in Section
2.2.3.5. It is important to notice that single aerobic
tests can only be performed when the tests can be
executed immediately after collection (and preferably
avoiding a washing procedure).
b. After the activated sludge has been transferred, keep
the same redox conditions as those prevailing in the
collection tank (e.g. anaerobic or anoxic) as described
in Section 2.2.3.5 for at least 30 min while confirming
that the pH and temperature are at the target values of
interest (otherwise adjust and wait until stable
conditions are reached). Do not start to aerate the
sample and, if available, sparge N2 gas (or another gas
available) to avoid oxygen intrusion.
c. Repeat steps 'b' to 'e' from Test EBPR.ANA.1.
d. The test starts at 'time zero' with the supply of air (or
pure oxygen), ensuring that the DO concentration
with regard to the DO saturation concentration at
local conditions reaches at least 2.0 mg DO L-1 within
the first 10 min of execution of the aerobic test and
around 4-5 mg DO L-1 onwards.
e. After the air supply starts, keep track of the execution
and sampling time with a stopwatch.
f. Duration and sampling of the aerobic stage:
i. The aerobic EBPR test can last between 2 - 4 h.
ii. The DO concentration in the bulk liquid can be
monitored throughout the test with the use of a
DO probe.
iii. For the estimation of the aerobic kinetic
parameters, samples for the determination of PO4
must be collected every 5 min in the first 30-40
min of execution of the test. After this period, the
sampling frequency can be reduced to every 10 or
36
15 min during the first 1 h, and later on to every
15 or 30 min until the test is finished.
iv. If the aerobic kinetic conversions of PHAs and
glycogen are of interest, samples can be collected
at the same time as the samples for PO4
determination.
v. To estimate the aerobic stochiometric
conversions, samples for the analytical
determination of total P, PO4, and MLSS and
MLVSS concentrations, as well as for PHAs and
glycogen (if applicable), must be collected both
at the start and at the end of the aerobic phase.
However, the total oxygen consumption must be
determined by respirometry (as presented in
Chapter 3).
g. Conclude the aerobic EBPR batch test with the
collection of the last samples needed for the
estimation of the aerobic stoichiometric conversions.
h. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
Test EBPR.AER.2 Combined anaerobic-aerobic EBPR batch tests
a. Execute the anaerobic test as follows:
i. Repeat steps 'a' to 'g' from Test EBPR.ANA.1.
ii. Afterwards, execute steps 'b' to 'e' from Test
EBPR.ANA.2.
b. Immediately after the execution of the anaerobic test,
the aerobic EBPR batch test can start with the
sparging of compressed air or pure oxygen ensuring
that the DO concentration with regard to the DO
saturation concentration at local conditions reaches at
least 2.0 mg DO L-1 within the first 10 min of
execution of the aerobic test and around 4-5 mg DO
L-1 afterwards.
c. Repeat step 'f' from Test EBPR.AER.1.
d. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
Test EBPR.AER.3 Combined anaerobic-anoxic-aerobic EBPR
batch tests in series
a. Execute the anaerobic test as follows:
i.
Repeat steps 'a' to 'g' from Test EBPR.ANA.1.
ii.
Afterwards, execute steps 'b' to 'e' from Test
EBPR.ANA.2.
b. Immediately continue with the execution of the
anoxic test by repeating steps 'd' and 'e' from Test
EBPR.ANOX.1.
c. After the anoxic test, continue with the execution of
the aerobic stage by repeating steps 'd' to 'g' from Test
EBPR.AER.1.
d. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Test EBPR.AER.4 Combined anaerobic-anoxic-aerobic EBPR
batch tests in parallel
a. This test requires two (preferably identical) reactors:
one for the execution of an anaerobic-anoxic stage
and another for the execution of a single aerobic test.
This is achieved by transferring, at the end of the
anaerobic test, a defined volume (usually 50 %) from
the bioreactor where the anaerobic-anoxic test will
take place to the second bioreactor where the aerobic
test will be conducted. For this purpose, the steps
described here below are recommended.
b. Execute the anaerobic test as follows:
i.
Repeat steps 'a' to 'g' from Test EBPR.ANA.1.
ii.
Afterwards, execute steps 'b' to 'e' from Test
EBPR.ANA.2.
c. Once the anaerobic stage is completed, transfer 50 %
of the activated sludge present in the anaerobic
bioreactor to the empty bioreactor (aerobic
bioreactor).
d. Continue with the execution of the anoxic test in the
same bioreactor where the anaerobic test took place
by repeating steps 'd' and 'e' from Test
EBPR.ANOX.1.
e. In parallel, carry out the execution of the aerobic test
in the second bioreactor (where sludge was
transferred) by repeating steps 'd' to 'g' from Test
EBPR.AER.1. Note that the execution of two (anoxic
and aerobic) tests in parallel requires more advanced
experimentation skills. Two alternative approaches
are that either two persons execute the tests, or to
carry out the tests one after another.
f. Repeat steps 'j' to 'n' from Test EBPR.ANA.1.
2.2.5 Data analysis
2.2.5.1 Estimation of stoichiometric parameters
Prior to estimation of the stoichiometric and kinetic
parameters, a COD balance should be conducted to
validate the results and confirm their quality and
reliability (Barker and Dold, 1995). A rather important
tool to assess the reliability of data is the COD balance.
Theoretically, the COD must be conserved in a truly
anaerobic system so that the total COD entering an
anaerobic stage must be equal to the total COD that
leaves the anaerobic stage (Wentzel et al., 2008). Thus,
assuming that soluble COD, intracellular PHAs and
glycogen are the only carbon components involved in the
chemical transformations, a COD balance can be
performed as follows:
ACTIVATED SLUDGE ACTIVITY TESTS
37
CODB,cons + CODGLY,cons = CODPHA,prod
Eq. 2.2.1
Where:
CODB,cons is the concentration of biodegradable substrate,
as COD, consumed during the duration of the
anaerobic batch activity test, in mg COD L-1.
CODGLY.cons is the concentration of intracellular glycogen
consumed during the duration of the anaerobic batch
activity test, in mg COD L-1.
CODPHA,prod is the concentration of intracellular PHAs
formed or stored during the duration of the anaerobic
batch activity test, in mg COD L-1.
Also, the percentage of error to close the COD
balance (ΔCOD (%)) can be estimated as:
The percentage of error to close the COD balance
(ΔCOD (%)) can be estimated as:
% COD balance = [1 ‒
CODB,cons + CODGLY,cons ‒ CODPHA,prod
CODB,cons + CODGLY,cons + CODPHA,prod
Where:
CODPHA,cons is the total concentration of PHA
consumption during the aerobic batch activity test, in
mg COD L-1.
CODGLY,prod is the total concentration of glycogen
produced during the aerobic batch activity test, in mg
COD L-1.
CODBio,prod is the total concentration of biomass produced
during the duration of the aerobic batch activity test,
in mg COD L-1.
ΔO2,cons is the total concentration of oxygen consumed in
the aerobic batch activity test estimated based on
respirometry and oxygen uptake rates (see Chapter 3
on Respirometry) in mg COD L-1.
]·100
Eq. 2.2.2
Ideally, ΔCOD (%) should be lower than 1-5 % but
values as high as 15 % are often reported and considered
acceptable in view of, and depending on, data quality and
the uncertainty that the determination of certain
parameters creates (in particular glycogen). The reader is
invited to consult Chapter 5 for more information on the
assessment of data quality.
Similarly to the anaerobic transformations, the COD
balance can be an important tool to assess the reliability
of the data obtained in anoxic and aerobic batch activity
tests (Ekama and Wentzel, 2008a,b). Thus, the total
amount of final electron acceptors consumed should
equal the total amount of electron donors oxidized. For
the aerobic transformations that take place during an
aerobic EBPR test, a COD balance can be performed as
follows:
∆CODcons = CODinput ‒ CODoutput = ∆O2,cons
Eq. 2.2.3
Assuming that PHAs, glycogen (GLY) and biomass
(Bio) are the only COD components that change during
the anoxic or aerobic phase, the net COD consumption
can be calculated as follows:
CODPHA,cons ‒ CODGLY,prod ‒ CODBio,prod = ∆O2,cons
Eq. 2.2.4
% COD balance = [1 ‒
CODPHA,cons ‒ CODGLY,prod ‒ CODBio,prod ‒ ∆O2,cons
] · 100
CODPHA,cons + CODGLY,prod + CODBio,prod + ∆O2,cons
Eq. 2.2.5
Ideally, similar to the determination of COD balances
for the anaerobic stage of EBPR batch activity tests,
ΔCOD(%) should be lower than 1-5 % but values as high
as 10 % can be considered acceptable. Chapter 5 provides
different tools and approaches to assess the quality of the
data obtained and execute the estimation of the COD
balances with a smaller uncertainty about their reliability.
COD balances can also be determined and applied to
anoxic EBPR activity tests where, for instance, nitrate or
nitrite act as the final electron acceptor. When applying
a COD balance on EBPR batch activity tests executed
with a different final electron acceptor rather than
oxygen, the approach would be similar, but: (i) the
equivalent COD concentrations of the final electron
acceptor should be determined and expressed in COD
units based on its capacity to accept electrons, and (ii)
the corresponding maximum stoichiometric yields (Y)
of the metabolic conversions on the final electron
acceptor of interest (for instance, nitrate or nitrite)
should be known. For tests conducted with nitrate as the
final electron acceptor, the maximum stoichiometric
yields estimated by Kuba et al. (1996) can be used. Last
but not least, the total consumption of the final electron
acceptor during the execution of the anoxic batch EBPR
test should be estimated based on the experimental
methods described in Chapter 3.
38
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Once the quality and reliability of the data have been
confirmed and validated, the stoichiometric and kinetic
parameters can be computed. Often, Net P-released,
glycogen conversion, and PHAs produced per organic
carbon or COD consumed (YVFA_PO4,An, YGly/VFA,An,
YVFA_PHA,An, YVFA_PHB,An and YVFA_PHV,An ratios,
respectively) are the stoichiometric parameters of interest
to assess the anaerobic stoichiometry of the EBPR
processes (Table 2.2.1).
release occurs continuously throughout the anaerobic test
as a consequence of the anaerobic maintenance
requirements of the cells (though usually it can only be
observed after the carbon source is depleted), the
accumulated PO4 released due to anaerobic maintenance
should be excluded from the total PO4 release observed.
Thus, the Net Preleased can be estimated as follows:
Net Preleased = (Total PO4 -Preleased ) ‒
Regarding the anaerobic stoichiometric parameters,
often the most common carbon source used for the
execution of EBPR batch activity tests is acetate (Ac).
Figure 2.2.5 shows a graphic representation of the
determination of both the stoichiometry and kinetic
parameters for Ac consumed and orthophosphate
released in an anaerobic batch activity test. It should be
noted that the Net Preleased due to Ac uptake should be
estimated after excluding the secondary PO4 release
(rPP_PO4,Sec,An). However, because the secondary PO4
(rPP_PO4,Sec,An) · (test duration)
Where:
Net Preleased is the PO4 released due to Ac uptake only, mg
PO4-P L-1.
rPP_PO4,Sec,An corresponds to the PO4 release rate due to
anaerobic maintenance requirements of the biomass,
mg PO4-P L-1 h-1.
PO4 released due
to maintenance
rPP_PO4,Sec,An
(mg PO4-P L-1 h-1)
HAc, PO4 (mg P L-1)
PO4
Total PO4 released
(mg PO4-P L-1)
HAc consumed
(mg Acetate L-1)
Net P-released = PO4 released {(rPP_PO4,Sec,An) . (duration of test [h])}
rAc,An
(mg Acetate L-1 h-1)
rPP_PO4,An
(mg PO4-P L-1 h-1)
Eq. 2.2.6
HAc
PO4 released due
to maintenance
Time (h)
Figure 2.2.5 Example of the determination of the maximum anaerobic volumetric kinetic rates for acetate consumption (Ac) and orthophosphate released
(PO4) in an anaerobic batch activity test where all the carbon is consumed. For the estimation of the Net P released, the secondary P release
(rPP_PO4,Sec,An)(corresponding to the anaerobic endogenous maintenance requirements which occur throughout the anaerobic test) must be excluded (Net P
released = [(Total PO4 released)] - [(rPP_PO4,Sec,An)·(duration of the test)]). The Net P released should be used for the determination of the anaerobic PO4-to-Ac
stoichiometric ratio (YAc_PO4,An).
The net P released should be used for the
determination of the anaerobic PO4/Ac stoichiometric
ratio (YAc_PO4,An). Thus, for instance, if in a batch activity
test the carbon is fully consumed, the anaerobic net Preleased/Ac-consumed ratio (which can also be referred
to as the P/C ratio) of a given culture can be calculated as
follows:
YAcPO4 ,An =
SPO4 ,ini ‒ SPO4 ,final
Net Preleased
=
SAc,ini ‒ SAc,final
SAc,cons
Eq. 2.2.7
Where:
SAc,cons is the concentration of acetate consumed in the
batch activity test, mg L-1.
ACTIVATED SLUDGE ACTIVITY TESTS
SPO4,ini is the orthophosphate concentration in the bulk
liquid at the beginning of the batch activity test, mg
PO4-P L-1.
SPO4-P,final is the orthophosphate concentration in the bulk
liquid at the time when the acetate concentration is
consumed or at the end of the anaerobic batch activity
tests if not all the acetate is depleted, mg PO4-P L-1.
SAc,ini is the concentration of acetate in the bulk liquid at
the beginning of the batch activity test, mg L-1.
SAc,final is the concentration of acetate in the bulk liquid at
the end of the batch activity test, mg L-1.
However, if the carbon is not depleted (e.g. when
carbon is added in excess or the duration of the test is
relatively too short to allow the full consumption of the
carbon source), then the difference between the initial
and final concentration of the compounds can be divided
by the difference between the initial and final
concentrations of carbon (net carbon consumed). A
similar approach can be applied for the determination of
other anaerobic stoichiometric ratios of interest (like the
anaerobic glycogen hydrolysis and PHAs produced per
C-source consumed ratios). Often, different units are
used to report the stoichiometry of the anaerobic
conversions, like the use of C-mol or P-mol instead of mg
Ac or mg PO4-P in scientific publications. For this
purpose, Appendix I contains a series of coefficients for
the conversion of units of certain compounds of interest
(e.g. to convert the units of orthophosphate from g PO4P to P-mmol). Unlike the anaerobic stoichiometric
parameters, the anoxic and aerobic stoichiometric
parameters cannot be determined in a similar
straightforward manner. Because EBPR cultures have
different intracellular metabolic conversions occurring
simultaneously under anoxic or aerobic conditions (polyP uptake, glycogen replenishment, growth and
maintenance) (Table 2.2.1), the net consumption of
electron donors (PHAs) and final electron acceptors
(oxygen, nitrate or nitrite) is the combined result of these
four overlapping metabolic activities (Smolders et al.,
1994b).
39
Nevertheless, Smolders et al. (1994b) and Kuba et al.
(1993, 1996) proved that using metabolic modelling, the
four aerobic or anoxic metabolic processes are dependent
on the ratio between the ATP produced per NADH
consumed during aerobic or anoxic respiration, the so
called ‘δ-ratio’ or ‘δ-value’ (Smolders et al., 1994b;
Kuba et al., 1996). This means that for EBPR cultures,
the aerobic or anoxic value of δ can be determined to
consequently estimate the values of the different aerobic
or anoxic stochiometric parameters, respectively.
Furthermore, despite different δ values having been
reported in literature for diverse microbial populations
(ranging from 1.3 to 2.2) (Lopez-Vazquez et al., 2009a),
for EBPR cultures, the effect of the δ-value is rather
insensitive to the aerobic and anoxic stoichiometric ratios
within this range of δ values. The latter suggests that the
aerobic and anoxic stoichiometric parameters are also
insensitive and there may not be any need for their
determination if the plant or lab-scale system operates
under regular operating and environmental conditions.
However, if the determination of the δ value is needed,
two tests can be performed with and without the presence
of orthophosphate in the bulk liquid and by measuring the
oxygen uptake rate (OUR) by respirometry (Chapter 3).
By computing the differences in P-uptake rates and OUR
between the tests performed with and without the
presence of orthophosphate in the bulk liquid, the δ value
can be estimated as described by Smolders et al. (1994b).
Alternatively, mathematical modelling can be applied to
estimate the δ value based on the anoxic and/or aerobic
profiles of PHAs, glycogen, orthophosphate, growth rate
and maintenance requirements (Lopez-Vazquez et al.,
2009a). Should there be interest in the determination of δ
for EBPR cultures, the reader may need to refer to cited
references for further reading since this procedure falls
out of the scope of this book.
As a reference and guide, Table 2.2.3 shows different
parameters of interest for lab-scale EBPR cultures
enriched under different operating conditions (e.g.
carbon source) and dominated by either PAOs or GAOs.
ANAEROBIC
Table 2.2.3 Typical stoichiometric parameters of interest for lab-scale enriched EBPR cultures cultivated under standard conditions (20 °C, pH 7, 7-8 d SRT).
Stoichiometric parameter
Lab-scale PAO culture enriched with acetate
Anaerobic orthophosphate release to acetate uptake ratio
Anaerobic glycogen utilization to acetate uptake ratio
Anaerobic PHA formation to acetate uptake ratio
Anaerobic PHB formation to acetate uptake ratio
Anaerobic PHV formation to acetate uptake ratio
Anaerobic PH2MV formation to acetate uptake ratio
Common notation Units
YAc_PO4,An
YGly/Ac,An
YAc_PHA,An
YAc_PHB,An
YAc_PHV,An
YAc_PH2MV,An
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
Typical values Reference
0.50
0.50
1.22
1.10
0.12
N/A
Smolders et al. (1994a)
AEROBIC
ANAEROBIC
ANOXIC
ANAEROBIC
AEROBIC
ANAEROBIC
AEROBIC
40
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Aerobic poly-P formation to PHA consumption ratio
Aerobic glycogen formation to PHA consumption ratio
Aerobic PAO biomass growth to PHA consumption ratio
Aerobic maintenance rate of PAO
Aerobic poly-P formation to oxygen consumption ratio
Aerobic glycogen formation to oxygen consumption ratio of PAO
Aerobic PAO biomass growth to oxygen consumption ratio
Aerobic endogenous respiration rate of PAO
Lab-scale PAO culture enriched with propionate
Anaerobic orthophosphate release to propionate uptake ratio
Anaerobic glycogen utilization to propionate uptake ratio
Anaerobic PHA formation to propionate uptake ratio
Anaerobic PHB formation to propionate uptake ratio
Anaerobic PHV formation to propionate uptake ratio
Anaerobic PH2MV formation to propionate uptake ratio
Aerobic poly-P formation to PHA consumption ratio
Aerobic glycogen formation to PHA consumption ratio
Aerobic PAO biomass growth to PHA consumption ratio
Aerobic maintenance rate of PAO
Aerobic Poly-P formation to oxygen consumption ratio
Aerobic Glycogen formation to oxygen consumption ratio of PAO
Aerobic Biomass growth to oxygen consumption ratio
Aerobic endogenous respiration rate of PAO
Lab-scale DPAO culture enriched with acetate
Anaerobic orthophosphate release to acetate uptake ratio
Anaerobic glycogen utilization to acetate uptake ratio
Anaerobic PHA formation to acetate uptake ratio
Anaerobic PHB formation to acetate uptake ratio
Anaerobic PHV formation to acetate uptake ratio
Anaerobic PH2MV formation to acetate uptake ratio
Anoxic poly-P formation to PHA consumption ratio
Anoxic glycogen formation to PHA consumption ratio
Anoxic PAO biomass growth to PHA consumption ratio
Anoxic maintenance rate of PAO
Anoxic poly-P formation to NO3 consumption ratio
Anoxic glycogen formation to NO3 consumption ratio
Anoxic PAO biomass growth to NO3 consumption ratio
Anoxic endogenous respiration rate of PAO on NO3
Lab-scale GAO culture enriched with acetate
Anaerobic glycogen utilization to acetate uptake ratio
Anaerobic PHA formation to acetate uptake ratio
Anaerobic PHB formation to acetate uptake ratio
Anaerobic PHV formation to acetate uptake ratio
Anaerobic PH2MV formation to acetate uptake ratio
Aerobic glycogen formation to PHA consumption ratio
Aerobic GAO biomass growth to PHA consumption ratio
Aerobic maintenance rate of GAO
Aerobic glycogen formation to oxygen consumption ratio
Aerobic PHA degradation to oxygen consumption ratio of GAO
Aerobic endogenous respiration rate of GAO
YPHA_PP,Ox
YPHA_Gly,Ox
YPHA_PAO,Ox
mPAO,Ox
YPP
YGly,PAO
YPAO
mPAO,O2
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1 h-1
P-mol C-mol-1
C-mol mol-O2-1
C-mol mol-O2-1
mol-O2 C-mol-1 h-1
3.68
0.90
0.74
4x10-3
3.27
3.92
2.44
4.5x10-3
Smolders et al. (1994b)
YPr_PO4,An
YPr_Gly,An
YPr_PHA,An
YPr_PHB,An
YPr_PHV,An
YPr_PH2MV,An
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
0.42
0.32
1.23
0.04
0.55
0.65
Oehmen et al. (2005c)
YPHA_PP,Ox
YPHA_Gly,Ox
YPHA_PAO,Ox
mPAO,Ox
YPP
YGly,PAO
YPAO
mPAO,O2
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1 h-1
P-mol C-mol-1
C-mol mol-O2-1
C-mol mol-O2-1
mol-O2 C-mol-1 h-1
3.34
1.06
0.80
4x10-3
3.34
6.16
2.03
4.5x10-3
Oehmen et al. (2007)
YAc_PO4,An
YGly/Ac,An
YAc_PHA,An
YAc_PHB,An
YAc_PHV,An
YAc_PH2MV,An
YPHA_PP,Ax
YPHA_Gly,Ax
YPHA_PAO,Ax
mPAO,Ax
YNO3_PP,Ax
YNO3_Gly,Ax
YNO3_PAO,Ax
mPAO,NO3
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1 h-1
P-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
N-mol C-mol-1 h-1
0.50
0.50
1.22
1.10
0.12
N/A
0.46
1.27
1.63
3.64x10-3
0.414
0.35
0.57
3.27x10-3
Smolders et al. (1994a),
Kuba et al. (1996)
YGly/Ac,An
YAc_PHA,An
YAc_PHB,An
YAc_PHV,An
YAc_PH2MV,An
YPHA_Gly,Ox
YPHA_GAO,Ox
mGAO,Ox
YGly
YPHA,GAO
mGAO,O2
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1 h-1
C-mol mol-O2-1
C-mol mol-O2-1
mol-O2 C-mol-1 h-1
1.12
1.86
1.36
0.46
0.04
0.95
0.75
3.06x10-3
4.89
2.18
3.51x10-3
Zeng et al. (2003a)
Kuba et al. (1996)
Zeng et al. (2003a)
ACTIVATED SLUDGE ACTIVITY TESTS
Regarding the anaerobic kinetics, the most important
parameters are the maximum specific consumption rate
of carbon source or VFA (qVFA,An), the maximum specific
P-release rate (qPP_PO4,An), PHA formation rate
(qVFA_PHA,An) and the endogenous ATP maintenance
coefficient (mATP,An) (Table 2.2.1). They can be
computed by plotting the experimental data (y-axis)
versus time (x-axis) and fitting the experimental data
obtained in the anaerobic batch activity tests using linear
regression. Because one is interested in the maximum
rates, a linear regression approach can be applied by
fitting the very first set or group of experimental data
points obtained at the beginning of the batch activity test.
This is the main reason why the sampling frequency
within the first 30-40 min of execution of the batch
activity tests is set to 5 min. Preferably the linear
regression must be carried out by plotting the
concentrations and fitting more than 4-5 experimental
data points by linear regression while achieving a
statistical determination coefficient (R2) not lower than
0.90-0.95. With the estimation of the linear regression
equation (of the form: ‘y = Ax + B’), the maximum
volumetric kinetic rates of the parameters of interest can
be determined with the ‘A’ coefficient of the linear
regression expression which corresponds to the ‘slope’ of
the set of data points. This will result in the determination
of the maximum volumetric rates (usually reported in
units such as mg L-1 h-1 or g m-3 d-1). Figure 2.2.5
illustrates the estimation of the maximum anaerobic
kinetic rates for an EBPR culture. The reader is referred
to Chapter 5 for a more detailed description of other
alternative and advanced statistical methods and tools for
the determination of the rates. For activated sludge
samples from full-scale systems, the rates can be
expressed as maximum specific kinetic rates by dividing
the volumetric rates (or values of the slopes) by the
concentration of activated sludge volatile suspended
solids (VSS). However, due to the particular dynamic
behaviour of the intracellular compounds present in
EBPR cultures, often the maximum specific kinetic rates
are expressed in terms of the active biomass fraction
(excluding the presence of intracellular compounds)
quantified according to the following approximate
expression:
Active biomass fraction = MLVSS ‒ PHA ‒ glycogen
Eq. 2.2.8
From a microbiological perspective, the previous
equation is not an accurate expression due to the potential
accumulation of non-biodegradable organics and,
logically, the intrinsic metabolic activity of every single
cell. Nevertheless, it is a commonly accepted estimate for
experimental purposes in the wastewater treatment
practice. Similarly to carbon and phosphate compounds
(particularly for lab-cultivated cultures), the active
biomass fraction can also be expressed in C-mol units
instead of mg VSS. For this purpose the elemental
composition of the biomass is applied (see Appendix I
with the corresponding conversion unit factors for
different compounds). Commonly, once the maximum
specific kinetic rates are found they can be reported for
the compound of interest as mg g VSS-1 h-1 or mg g VSS-1
d-1. Similarly, the anaerobic endogenous ATP
maintenance coefficient (mATP,An) can be determined
based on the profile of orthophosphate concentration
recorded during the execution of an anaerobic EBPR
batch activity test performed under the absence of any
carbon source (Section 2.2.4.1). This maximum
volumetric rate resembles or corresponds to the
endogenous P-release rate (mPP_PO4,Sec,An) observed in
anaerobic tests once the carbon source is depleted (Fig.
2.2.6). Then, mATP,An is equivalent to mPP_PO4,Sec,An
(Wentzel et al., 1989a; Smolders et al., 1994a). A similar
linear regression approach can be used for the
determination of the maximum specific kinetic rates for
the aerobic and anoxic batch activity tests (tables 2.2.2
and 2.2.3). Usually, orthophosphate uptake, biomass
growth, PHA degradation, glycogen formation and
aerobic maintenance requirements are the kinetic
parameters of interest dependent on the final electron
acceptor available. Figure 2.2.6 displays an example to
illustrate the determination of the maximum kinetic rates
in an aerobic (or anoxic) batch activity test.
rPO4_PP,Ox
(mgPO4-P L-1 h-1)
PO4-P, NH4-N (mg L-1)
2.2.5.2 Estimation of kinetic parameters
41
or
(if anoxic conditions)
rPO4_PP,Ax
(mgPO4-P L-1 h-1)
rNH4_Bio,Ox
(mg NH4-N L-1 h-1)
Time (h)
Figure 2.2.6 Example of the determination of the maximum aerobic (or
anoxic) volumetric kinetic rates for orthophosphate uptake (PO4) and
ammonia consumption (NH4) in an aerobic (or anoxic) batch test.
42
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
It is important to mention that the maximum specific
biomass growth rate (qPHA_Bio,Ox) cannot be computed
straightforwardly by following the increase in biomass
concentrations during the cycle. Instead, the maximum
specific ammonia (NH4) consumption rate (qNH4_Bio,Ox)
divided by the nitrogen composition of the active
biomass fraction (0.20 N-mol per C-mol biomass)
(Smolders et al., 1995; Zeng et al., 2003a; LopezVazquez et al., 2007; Welles et al., 2014) can be used.
However, the latter approach may be valid as long as
nitrification is absent, no chemical precipitation or
adsorption of NH4 occurs and the relatively small
differences allow a satisfactory determination of NH4.
Often, differences in NH4 concentrations may be
negligible or fall into the standard error of the analytical
technique. This may complicate the determination
process.
The experiments presented in this chapter focus on
the relative conversions of intracellular and soluble
compounds present in the water phase. It does not cover
or tackle the consumption profiles nor the analysis of
final electron acceptors (such as oxygen consumption,
oxygen uptake rates, or nitrate uptake rates). These
parameters are presented and discussed in Chapter 3
dealing with respirometry.
The determination of the aerobic maintenance
coefficient of EBPR cultures is of major importance. For
the determination of this parameter, extended aeration
tests (regularly of at least 24 h) must be executed under
the absence of an external carbon source. After 24 h, the
aerobic ATP requirements (mATP,Ox) can be determined
based on the oxygen consumption or uptake rate (OUR)
(as O2 mol consumed per mol active biomass per hour)
and the aerobic maintenance requirements be computed
with the following expression as a function of δ (with an
average typical value of around 1.75-1.80 for enriched
EBPR cultures):
mO2 =
1.125 mATP,Ox
2.25δ + 0.5
Eq. 2.2.9
2.2.6 Data discussion and interpretation
The results from the EBPR batch activity tests will
provide important information regarding not only the
biomass activity under different operating and
environmental conditions but also about the general state
of the biomass and some directions regarding the
dominant microbial populations present in the sludge
sample. Concerning the latter, the reader is invited to
consult chapters 7 and 8. Furthermore, the identification
of the dominant microbial species is an important
complement to a better understanding of the EBPR
process activities.
2.2.6.1 Anaerobic batch activity tests
As PAOs are the only organisms known to release
phosphate during the anaerobic uptake of VFA, the
anaerobic P-released/C-uptake or P-released/COD ratios,
YC_PO4,An such as YVFA_PO4,An, YAc_PO4,An or YPr_PO4,An, may
be considered one of the most suitable indicators to
assess the PAO and GAO activity of EBPR cultures. A
high P-released/C-uptake ratio has been considered as an
indicator for PAO presence, whereas a low ratio has been
considered to indicate the significant presence of GAOs.
Contradictorily, such a consideration has been the subject
of controversy (Schuler and Jenkins, 2003; Oehmen et
al., 2007; Lopez-Vazquez et al., 2008b; Welles et al.,
2015b). Such a controversy arises due to the broad range
of ratios reported in literature, ranging from values as low
as 0.025 to even 0.75 P-mol C-mol-1 or higher (Schuler
and Jenkins, 2003). Several studies with highly enriched
PAO cultures revealed that factors other than the
presence of GAOs affect the anaerobic P released/C
uptake as well. For instance, the carbon source, the pH
and the poly-P content of the PAOs and the specific
clades of PAOs enriched in the study have been shown to
affect the P-released/C-uptake ratio (Smolders et al.,
1994a; Filipe et al., 2001a; Zhou et al., 2008; Acevedo et
al., 2012). Consequently, it is not advisable to use only
this ratio as the direct and sole indicator to assess the
EBPR activity of an activated sludge. Nevertheless, it
may well be used to provide a rough estimation of the
dominant metabolisms prevailing in the activated sludge.
If supported by molecular techniques (chapters 7 and 8
from this book), it can offer an adequate and more
complete overview of the biomass activity. Thus, based
on observations drawn from past research, different P/C
ratios performed under standard conditions (20 °C, pH
7.0) and fed with acetate may indicate (Schuler and
Jenkins, 2003):
P/C ratio (P-mol C-mol-1)
< 0.25
0.25-0.50
> 0.50
Dominant metabolism
GAO-dominated metabolism
Intermediate PAO-GAO metabolism
PAO-dominated metabolism
Figure 2.2.7 shows different anaerobic P-released/Cuptake ratios reported in literature as a function of the
amount of phosphorus accumulated in the sludge on a
TSS concentration basis expressed as the P/TSS ratio of
ACTIVATED SLUDGE ACTIVITY TESTS
43
the sludge. Interestingly, the phosphorus content of the
sludge (in terms of mg P g VSS-1) has been shown to have
a strong correlation with the anaerobic P-released/Cuptake ratio. As observed in Figure 2.2.7, an anaerobic
P/C ratio higher than 0.50 is usually observed when the
P/TSS content of the enriched PAO sludge is higher than
0.10 g P g TSS-1. Thus, probably an activated sludge
system will have a satisfactory EBPR activity when the
anaerobic P-released/C-uptake and sludge P/TSS ratios
are around or higher than 0.50 P-mol C-mol-1 and 0.10
mg P mg TSS-1, respectively. Lower values could suggest
that the EBPR activity of the system may be limited by
certain operational or environmental factors such as (i)
the relatively abundant presence of GAOs instead of
PAOs, (ii) intrusion of electron acceptors into the
anaerobic phase (like oxygen, NO3, NO2), (iii) the
addition of Al of Fe salts for chemical P-removal, or
occasionally (iv) the presence of inhibitory or toxic
compounds. Under such circumstances the activated
sludge system must be carefully revised to implement the
corresponding corrective measures.
Anaerobic P-release/C-uptake ratio
(P-mol C-mol -1)
0.8
y = 4.53 x - 0.0153
R2 = 0.92
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.00 0.02
0.04
0.06
0.08 0.10 0.12
0.14
0.16 0.18
P/TSS (mg mg-1)
Figure 2.2.7 The anaerobic P-released/C-uptake ratio as a function of the
P/TSS ratio of the biomass as reported in literature (adapted from Schuler
and Jenkins, 2003; and Welles et al., 2015b).
Under certain circumstances, the EBPR activity tests
may be conducted under non-standard conditions (pH
different to 7.0) using real wastewater or other carbon
sources rather than acetate. For comparison and
benchmarking purposes, the P/C ratio may be corrected
using the expressions developed by Smolders et al.
(1994a) or Filipe et al. (2001a) respectively, for acetatefed cultures:
YAc_PO4,An = 0.19 · pH ‒ 0.85
Eq. 2.2.10
YAc_PO4,An = 0.16 · pH ‒ 0.55
Eq. 2.2.11
Also, the use of real wastewater or other C sources
rather than acetate (like propionate, butyrate or glucose)
will lead to a lower anaerobic P/C ratio than those
reported for EBPR cultures where the PAO metabolism
(or also PAOs) is dominant. Such lower P/C ratios can
be, besides a low P/TSS content as previously discussed,
a consequence of the lower energy needed for PHA
storage or due to a higher involvement of GAO
metabolism (possibly caused by the presence of GAOs).
If the glycogen consumption/C-uptake ratio does not
increase and remains within the commonly reported ratio
for PAO-dominated systems of 0.35-0.50 C-mol C-mol-1
(Smolders et al., 1994b; Schuler and Jenkins, 2003), the
system can be assumed to be robust and stable,
particularly if the intracellular P/TSS ratio is not
considerably lower than 0.10 in enriched cultures.
However, if under such circumstances the glycogen
consumption/C-uptake ratio, YGLY/Ac,An, increases above
0.35-0.50 C-mol C-mol-1, then the GAO metabolism (or
the presence of GAOs themselves) may start to dominate
the system and ultimately lead to the process
deterioration as long as this ratio continues to increase
(ultimately being able to reach a YGLY/Ac,An of 1.12 C-mol
C-mol-1) in combination with a decrease in the P/TSS
content below 0.10 mg P mg TSS-1.
In parallel, PHA synthesis/C-uptake ratios higher
than 1.33 C-mol C-mol-1 will be observed as a
consequence of a higher synthesis of PHV and PH2MV.
PHV/C-uptake ratios higher than 0.10 and up to 0.250.30 may be observed together with the formation of
PH2MV. These values probably indicate the potential
deterioration of the EBPR activity and efficiency.
Regarding the anaerobic kinetic rates, in particular the
initial maximum anaerobic uptake rate of carbon sources,
different values have been reported for PAO and GAO
lab-scale dominated cultures (mostly sequencing batch
reactors, SBR) and full-scale systems such as modified
UCT (University of Cape Town), Phoredox (which
stands for phosphorus reduction oxidation), and PhoStrip
(from phosphorus stripping) (Wentzel et al., 2008)
(Table 2.2.4).
44
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 2.2.4 Maximum initial specific anaerobic C-source uptake rates reported in literature for lab-scale and full-scale EBPR systems.
Lab-scale EBPR systems
Dominant microorganism/metabolism and system
PAOs - SBR
GAOs - SBR
Full-scale EBPR systems
Dominant microorganism/metabolism
PAOs - Modified UCT
PAOs - Phoredox
PAOs - Sidestream PhoStrip
qVFA,An in C-mol C-mol-1 h-1
0.27
0.20
0.20
0.20
0.20
0.24
0.16-0.18
0.20
0.19
Reference
Smolders et al. (1994a)
Filipe et al. (2001b)
Kuba et al. (1996)
Brdjanovic et al. (1997)
Lopez-Vazquez et al. (2007)
Filipe et al. (2001a)
Zeng et al. (2003a,b)
Lopez-Vazquez et al. (2007)
Lopez-Vazquez et al. (2009a)
qVFA,An in mg Ac g VSS-1 h-1
22
19
47
7-31
14
21
11
14
9
23
Reference
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Kuba et al. (1997a,b)
Kuba et al. (1997b)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Brdjanovic et al. (2000)
Similarly to the P-released/C-uptake ratio, the kinetic
rates seem to be dependent on the poly-P content of the
sludge (Schuler and Jenkins 2003, Welles et al., 2016,
submitted), ranging from 0.02 to 0.20 C-mol C-mol-1 h-1.
However, in most studies conducted with medium polyP contents and P/Ac ratios around 0.5 P-mol C-mol-1, the
observed uptake rates converge around 0.20 C-mol Cmol-1 h-1.
Similarly, rates observed in the tests performed with
activated sludge samples from full-scale EBPR systems
lie in between 17 and 22 mg Ac g VSS-1 h-1 (LopezVazquez et al., 2008a).
Temperature plays a major role in the different EBPR
microbial processes (Brdjanovic et al., 1997, 1998c,
Table 2.2.5).
Table 2.2.5 Arrhenius temperature coefficients (θ) reported in literature to describe the maximum specific kinetic rates of the different EBPR metabolic
processes occurring in lab-scale and full-scale EBPR systems (Meijer, 2004).
Parameter
qVFA,An
mATP,An
qPHA,Ox
qPHA_Gly,Ox
qPO4_PP,Ox
qPAO,Ox
mATP,Ax
mATP,Ox
θ*
1.094 (e0.090)
1.071 (e0.069)
1.129 (e0.121)
1.125 (e0.118)
1.031 (e0.031)
1.081 (e0.078)
1.071 (e0.090)
1.071 (e0.090)
Reference
Brdjanovic et al. (1998c); Meijer (2004)
Smolders et al. (1995); Murnleitner et al. (1997)
Brdjanovic et al. (1998); Meijer (2004)
Meijer (2004)
Murnleitner et al. (1997); Brdjanovic et al. (1998c)
Brdjanovic et al. (1997)
Murnleitner et al. (1997)
Murnleitner et al. (1997)
* The number in brackets displays the value of the Arrhenius temperature coefficient (θ) in terms of the Euler number.
ACTIVATED SLUDGE ACTIVITY TESTS
45
2.2.6.2 Aerobic batch activity tests
For comparison purposes, the observed maximum kinetic
rates must be standardized using Arrhenius coefficients.
Based on observations from Brdjanovic et al. (1997,
1998) and other authors (Smolders et al., 1995, and
Murnleitner et al., 1997), Meijer (2004), through the
development of the TUDelft model and the execution of
several modelling studies, defined suitable Arrhenius
temperature coefficients to best fit the maximum kinetic
rates of EBPR cultures for different operating lab- and
full-scale systems.
Although the aerobic kinetic rates reported in
literature for lab-scale EBPR systems are consistent
(Table 2.2.6), the values observed in full-scale EBPR
systems may vary widely (from modified UCT systems
to BIODENIPHO -biological denitrification nitrification
and phosphorus removal- plants) (Table 2.2.7).
Table 2.2.6 Maximum aerobic kinetic rates reported in literature for enriched EBPR cultures.
Culture or system
SBR
SBR
SBR
qPO4_PP,Ox
P-mol C-mol-1 h-1
0.055
0.046
0.083
qPHA_Gly, Ox
C-mol C-mol-1 h-1
0.080
-
qPAO,Ox
C-mol C-mol-1 h-1
0.014-0.016
0.13
-
mATP,Ox
mol ATP C-mol-1 h-1
1.9x10-3
1.2x10-3
1.7x10-3
Reference
Smolders et al. (1994b)
Brdjanovic et al. (1997)
Welles et al. (2014)
Table 2.2.7 Maximum initial specific aerobic P-uptake rates reported in literature for full-scale EBPR systems.
Dominant microorganism/metabolism
PAOs - Modified UCT
PAOs - Phoredox
PAOs - Sidestream PhoStrip
PAOs - Pilot-scale BIODENIPHO
Such considerable differences are logical since the
lab-scale systems are highly enriched with PAOs (based
on similar studies PAOs can probably compose more
than 80-90 % of the total populations) with little
variability in the percentage of enrichment (LopezVazquez et al., 2009a; Welles et al., 2014, 2015a),
whereas in full-scale systems PAOs may comprise
between 3 and 20 % of the total active biomass fraction
(Lopez-Vazquez et al., 2008a). Moreover, in full-scale
systems performing biological nitrogen and phosphorus
removal, EBPR cultures are exposed to alternating
anaerobic-anoxic-aerobic (A2O) stages that may reduce
the availability of intracellular PHAs after the sequential
qPO4_PP,Ox
mg P g VSS-1 h-1
19.2
9.0
13
4-6
8.0
9.1
6.2
6.3
9.8
2.2
4
Reference
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Kuba et al. (1997a, b)
Kuba et al. (1997b)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Brdjanovic et al. (2000)
Meinhold et al. (1999)
exposure to anoxic and aerobic conditions (since both
anoxic and aerobic metabolic activities require PHAs as
the carbon and energy source). Based on the data
provided in Table 2.2.7, it can be seen that values higher
than 10 mg P g VSS-1 h-1 are not common. Such kinetic
rates can be considered rather fast since, as previously
discussed, full-scale systems often tend to have aerobic
hydraulic retention times (HRT) of several hours (e.g. of
at least 6-8 h and longer) which will favour the uptake of
orthophosphate from the bulk liquid. Aerobic P-uptake
rates (qPO4_PP,Ox) lower than 10 mg P g VSS-1 h-1 will not
be nsatisfactory depending upon the length of the aerobic
stage (though extended aeration periods must be avoided)
46
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
(Brdjanovic et al., 1998c). Another important aspect to
consider is whether the plant is prone to chemical Pprecipitation because its occurrence will favour the
chemical P- removal to the detriment of the EBPR
process.
2.2.6.3 Anoxic batch activity tests
The simultaneous removal of orthophosphate and nitrate
(or nitrite) is highly desirable due to the potential savings
in energy and operational costs, while at the same time
keeping or maintaining a P-removal activity that can be
comparable to that observed under anaerobic-aerobic
conditions (Kuba et al., 1996). However, to a certain
extent, it has been a matter of controversy particularly
due to the rather variable and inconsistent anoxic Premoval activities observed in full-scale systems (Hu et
al., 2002) and theoretically decreased P-removal
potentially caused by lower PAO biomass yields. Table
2.2.8 shows an overview of different anoxic P-removal
activities observed in full-scale systems (anaerobicanoxic (A2) and anaerobic-anoxic-oxic (A2O) plants) that
are compared to the aerobic P-removal activities of the
same systems (following the protocols presented in this
chapter as introduced by Murnleitner et al., 1997).
Table 2.2.8 Maximum initial anoxic kinetic rates reported in literature for lab-enriched denitrifying EBPR cultures.
Culture or system
SBR, PAOs, A2 system
SBR, PAOs, A2 system
SBR, PAOs, A2 system
SBR, PAOs, A2O system
a
qPO4_PP,Ax
P-mol C-mol-1 h-1
0.1
0.02-0.63a
0.58a
0.33a
qPHA_Gly,Ax
C-mol C-mol-1 h-1
0.8
0.0025
0.9a
-
qPAO,Ax
C-mol C-mol-1 h-1
0.05
-
mAx
C-mol C-mol-1 h-1
3.6x10-3
-
Reference
Kuba et al. (1996)
Carvalho et al. (2007)
Zeng et al. (2003b)
Saito et al. (2004)
Units: P-mmol g VSS-1 h-1
Table 2.2.9 Maximum initial specific anoxic P-uptake rate and its relationship with maximum aerobic P-uptake rate reported for full-scale EBPR systems.
Dominant
microorganism/metabolism
PAOs - Modified UCT
PAOs - Phoredox
PAOs - Sidestream PhoStrip
PAOs - Pilot-scale BIODENIPHO
qPO4_PP,Ox
mg P g VSS-1 h-1
19.2
9.0
13
4-6
8.0
9.1
6.2
6.3
9.8
2.2
4
qPO4_PP,Ax
mg P g VSS-1 h-1
5.9
2.1
6
1.2-1.6
1.9
4.4
0.6
0.0
3.3
1.7
2
As observed in Table 2.2.9, the anoxic P-uptake rate
scarcely reaches more than 5 mg PO4-P g VSS-1 h-1 and,
even under certain circumstances, it is very low or absent.
In any case, the different anoxic P activities are a
combined reflection of (i) the level of enrichment of
denitrifying PAOs capable of using oxygen and nitrate
(and/or of other EBPR cultures and side-populations
involved in the denitrification process) (Kerr-Jespersen
and Henze, 1993; Meinhold et al., 1999; Saad et al.,
qPO4_PP,Ax / qPO4_PP,Ox
%
31 %
23 %
46 %
20-40 %
23 %
48 %
9%
0%
34 %
80 %
54 %
Reference
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Kuba et al. (1997a,b)
Kuba et al. (1997b)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Lopez-Vazquez et al. (2008a)
Brdjanovic et al. (2000)
Meinhold et al. (1999)
2016, submitted), and/or (ii) a measure of the level of
denitrifying capacity induced in PAOs (Kuba et al., 1996,
1997; Wachtmeister et al., 1997). In any case, the level
of exposure of the activated sludge system will favour the
growth of denitrifying EBPR populations and their
induction. Thus, higher anoxic P-uptake activities can be
expected in plant configurations operated with defined
pre-denitrification stage.
ACTIVATED SLUDGE ACTIVITY TESTS
2.2.7 Example
2.2.7.1 Description
To illustrate the execution of an EBPR batch activity test,
data from an anaerobic-aerobic test (Test EBPR.AER.2)
performed at 10 °C with a lab-enriched EBPR culture is
presented in this section. Test EBPR.AER.2 was carried
out to determine the anaerobic stoichiometry and the
anaerobic and aerobic kinetics of the EBPR processes.
Thus, the batch activity test was performed in a 2.5 L
bioreactor. All the equipment, apparatus and materials
were prepared as described in Section 2.2.3. pH and DO
sensors were calibrated less than 24 h before the test
execution. The test lasted 4.5 h and was composed of a
2.25 h anaerobic stage (created by continuously sparging
N2 gas throughout the test) followed by 2.25 h aerobic
stage (created by supplying compressed air in excess,
reaching a DO concentration higher than 4 mg L-1). Prior
to the batch test, 1.25 L of concentrated EBPR sludge
collected at the end of the aerobic phase of a lab-scale
bioreactor was transferred to the bioreactor and
acclimatized for 30 min at 10 °C under slow mixing (100
rpm) at pH 7.0 following the recommendations described
in Section 2.2.3.5. Activated sludge preparation for tests
was performed in less than 1 h after sludge collection.
Afterwards, 20 min before the start of the test, samples
for the determination of the parameters of interest were
collected (in accordance with the execution of Test
EBPR.AER.2).
The test started with the addition of 1.25 L synthetic
media containing 350 mg COD L-1 as Ac (other macroand micro-nutrients as well as 20 mg L-1 ATU were
included in synthetic media in accordance with Section
2.2.3.3). Because the test was executed at 10 °C, the
temperature of the synthetic media was adjusted to 10 °C
in a water bath operated at the same temperature before
addition. Because the main objective was to determine
the anaerobic stoichiometry and the anaerobic and
aerobic kinetics of the EBPR processes, samples were
collected more frequently (every 5 min) in the first 30
min of each anaerobic and aerobic phase. Immediately
47
after collection, all the samples were prepared, preserved
and stored prior to the analytical determination of the
parameters of interest (e.g. PO4, MLSS, MLVSS, and
PHAs, among other parameters) as described in Section
2.2.3.4 “Material preparation”. All the collected samples
were analysed as described in Section 2.2.2.5. In
particular, the intracellularly stored PHAs were
determined following the protocol for the determination
of PHB and PHV, and glycogen by the acid-hydrolysis
and extraction methods (Smolders et al., 1994a). Other
PHA compounds, like PH2MV, were not measured
because acetate was the carbon source supplied and
therefore it was expected that PHB and PHV would
comprise most of the PHAs. Table 2.2.10 shows the
experimental implementation plan of the execution of the
test.
2.2.7.2 Data analysis
Following up on the results from the experiment shown
in Table 2.2.10, Figure 2.2.8 shows the results from the
test displayed in the implementation plan and also an
estimation of the maximum volumetric kinetic rates by
applying linear regression. The different anaerobic and
aerobic conversions of the parameters of interest are
shown in Table 2.2.11, while Table 2.2.12 displays an
estimation of the different anaerobic and aerobic
stoichiometric and kinetic parameters of interest.
Overall, the results of the batch activity test (Figure
2.2.8) show the typical phenotype of a PAO-dominated
sludge: full Ac uptake in the anaerobic stage coupled
with anaerobic P release, PHA production and glycogen
consumption, while full PO4 uptake was observed in the
aerobic phase together with PHA utilization, glycogen
formation and slight NH4 consumption. Furthermore, the
relatively low VSS/TSS ratio observed at the beginning
and end of the test (of around 0.72-0.73) is typical for
EBPR systems due to poly-P accumulation (which is
reflected in a higher ash content) when compared to
systems that perform organic matter removal only (with
VSS/TSS ratios usually not lower than 0.80) (Wentzel et
al., 2008).
48
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 2.2.10 Example of an experimental implementation plan for the execution of a batch activity test (Type Test No. EBPR.AER.2) performed with a labenriched EBPR sludge at 10 °C using synthetic influent at pH 7.0.
Combined anaerobic-aerobic EBPR batch tests
Code: EBPR.AER.2
Date:
Description:
Test No.:
Duration
Thursday
17.12.2015 9:00 h
Experimental procedure in short:
Tests at 10 °C, pH 7, artificial substrate and enriched PAO culture 1. Confirm availability of sampling material and required equipment
3 of 6
2. Confirm calibration and functionality of the system, meters and sensors
4,5 h (270 min)
3. Transfer 1.25 L sludge to the batch reactor
Substrate:
Sampling point:
Samples No.:
Total sample volume:
Synthetic: Acetate (350 mg L ) + minerals
Middle mixed liquor height in the SBR
EBPR.AER.2(1-22)
305 mL
(10 mL for MLVSS, 12 mL for PHA, 4.5 mL for Glycogen,
6 mL for other samples)
2.5 L
-1
Reactor volume:
Sampling schedule
Time (min)
Time (h)
Sample No.
Parameter
-20
-0.33
1
-1
1
HAc (C-mmol L )
5.83
1
0
-1
PO 4-P (P-mmol L )
15
0.25
5
20
25
0.33
0.42
6
7
ANAEROBIC PHASE
30
0.50
8
40
0.67
9
50
0.83
10
4.85
4.57
3.98
3.48
2.87
2.21
1.35
0.43
0
0
0
0.24
0.45
1.01
1.35
1.69
2.14
2.85
3.01
3.04
3.07
3.11
1.26
20.04
12.68
1.39
20.08
12.71
-1
1.34
See table
140
2.33
13
145
2.42
14
150
2.50
15
3.00
2.73
2.45
155
2.58
16
160
165
2.67
2.75
17
18
AEROBIC PHASE
180
3.00
19
195
3.25
20
215
3.58
21
270
4.50
22
-1
HAc (C-mmol L )
-1
PO 4-P (P-mmol L )
-1
NH4-N (N-mmol L )
PHA (C-mmol)
Glycogen (C-mmol)
2.20
1.90
1.64
1.11
0.84
0
0
0
1.09
15.55
15.04
1.07
1.06
1.05
11.72
15.90
-1
MLSS and MLVSS (mg L )
See table
MLSS & MLVSS measurements
Sampling point
Cup No.
W1
W2
W3
W2-W1
W2-W3
MLSS
MLVSS
Ratio
Start anaerobic phase
1
2
3
0.08835
0.08835
0.08834
0.16525
0.16553
0.16435
0.10792
0.10997
0.10903
0.07690
0.07718
0.07601
End anaerobic/Start aerobic phase
4
5
6
0.08858
0.08848
0.08914
0.12437
0.12564
0.12527
0.09606
0.09646
0.09648
0.03579
0.03716
0.03613
End aerobic phase
7
8
9
0.08868
0.08764
0.08722
0.12859
0.12716
0.12622
0.09952
0.09881
0.09800
0.03991
0.03952
0.03900
0.05733
0.05556
0.05532
Average
0.02831
0.02918
0.02879
Average
0.02907
0.02835
0.02822
Average
7,690
7,718
7,601
7,670
3,579
3,716
3,613
3,636
3,991
3,952
3,900
3,948
5,733
5,556
5,532
5,607
2,831
2,918
2,879
2,876
2,907
2,835
2,822
2,855
0.75
0.72
0.73
0.73
0.79
0.79
0.80
0.79
0.73
0.72
0.72
0.72
Start Aner.
End Aner.
End Aer.
3,835
3,636
3,948
2,804
2,876
2,855
0.73
0.79
0.72
1,031
760
241.7
392.0
2
2
Sample taken before substrate addition
Biomass composition
Sampling point
-1
MLSS (mg L )
-1
MLVSS (mg L )
Ratio
-1
Ash (mg L )
-1
PHB (mg L )
-1
PHV (mg L )
-1
PHA (mg L )
-1
Glycogen (mg L )
%(PHA+Gly) MLVSS
-1
-1
Active biomass (mg L )
-1
Active biomass (Cmmol L )
90
1.50
12
135
2.25
13
See table
Average value of the concentration present in the synthetic substrate and and liquid phase of the sludge sample prior the start of the test
Sampling schedule (continued)
Time (min)
Time (h)
Sample No.
Parameter
60
1.00
11
08:40
08:40
08:50
09:00
09:05
11:15
13:30
13:45
14:00
10
0.17
4
1.32
12.27
15.09
MLSS and MLVSS (mg L )
4. Keep aerobic conditions with gentle mixing and air sparging at T and pH set
5. 20 min before starting, take sample for initial conditions (EBPR.AER.2(3.1))
6. Stop aeration and start sparging by N2 gas
7. Start cycle, add 1.25 L of sysnthetic media (0 min)
8. Continue sampling program according to schedule (5 min)
9. Stop sparging with N2 gas, start addition of air (135 min)
10. Stop sampling and aeration (after 270 min)
11. Organize the samples and clean the experimental equipment and space
12. Ensure all equipment is switched off and samples are handled properly
5
0.08
3
1
-1
NH4-N (N-mmol L )
PHA (C-mmol)
Glycogen (C-mmol)
1
0
0.00
2
Time (h:min)
08:00
08:10
08:20
Note:
Acetate (CH2 O)
30.03 mg C-mmol
-1
Ortho-Phosphate (PO4 -P)
31.00 mg P-mmol
-1
1,093
Ammonium (NH4 -N)
14.00 mg N-mmol
-1
232.1
PHB (CH1.5O 0.5 )
21.52 mg C-mmol
-1
20.9
37.3
18.6
PHV (CH1.6O 0.4 )
20.02 mg C-mmol
-1
262.6
429.3
250.8
Glycogen (CH10/6 O5/6 )
27.00 mg C-mmol
-1
423.7
343.3
429.2
Biomass (CH2.09 O0.54 N0.20)
26.00 mg C-mmol
-1
32.0
37.0
31.0
2,117
2,103
2,175
81.4
80.9
83.6
ACTIVATED SLUDGE ACTIVITY TESTS
49
Table 2.2.11 Summary of the anaerobic and aerobic conversions observed in the example of the batch activity test (Type Test No. EBPR.AER.2) performed
with a lab-enriched EBPR sludge at 10 °C using synthetic influent at pH 7.0.
Parameter
Ac
PO4-P
NH4-N
PHB
PHV
PHAs (PHB+PHV)
Glycogen
Unit
C-mmol L-1
P-mmol L-1
N-mmol L-1
C-mmol L-1
C-mmol L-1
C-mmol L-1
C-mmol L-1
Anaerobic phase
Start
[Time: 0]
5.20
0.00
1.32
11.23
1.05
12.27
15.69
End
[Time: 135 min]
0.00
3.11
1.39
18.21
1.86
20.08
12.71
Aerobic phase
Anaerobic
conversion
-5.20
3.11
0.07
6.99
0.82
7.80
-2.98
Start
[Time: 135 min]
0.00
3.11
1.39
18.21
1.86
20.08
12.71
End
[Time: 270 min]
0.00
0.00
1.05
10.79
0.93
11.72
15.90
Aerobic
conversion
0.00
-3.11
-0.34
-7.43
-0.93
-8.36
3.18
Table 2.2.12 Summary of the anaerobic and aerobic stoichiometric and kinetic parameters observed in the example of the batch activity test (Type Test No.
EBPR.AER.2) performed with a lab-enriched EBPR sludge at 10 °C using synthetic influent at pH 7.0.
Conversion
Anaerobic stoichiometry
Net P-released to Ac uptake ratioa
PHA production to Ac uptake ratio
PHV production to PHB production ratio
Glycogen consumption to Ac uptake ratio
Anaerobic kinetic ratesb
Maximum volumetric Ac uptake rate
Maximum specific Ac uptake rate
Maximum volumetric PO4 release rate
Maximum specific PO4 release rate
Secondary anaerobic PO4 release rate
Anaerobic maintenance coefficient
Aerobic kinetic rates
Maximum volumetric PO4 uptake rate
Maximum specific PO4 uptake rate
Volumetric aerobic NH4 consumption ratec
Maximum specific biomass growth rated
Symbol
Unit
Estimated value
YAc_PO4,An
YAc_PHA,An
YPHV/PHB,An
YGly/Ac,An
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
0.57
1.50
0.12
0.57
rAc,An
qAc,An
rPP_PO4,An
qPP_PO4,An
rPP_PO4,Sec,An
mPP_PO4,An
C-mmol L-1 h-1
C-mol C-mol -1 h-1
P-mmol L-1 h-1
P-mmol L-1 h-1
P-mmol L-1 h-1
P-mol C-mol-1 h-1
6.15
0.075
4.42
0.054
0.063
7.69 E-04
rPO4_PP,Ox
qPO4_PP,Ox
rNH4_Bio,Ox
qPAO,Ox
P-mmol L-1 h-1
P-mol C-mol-1 h-1
N-mol L-1 h-1
C-mol C-mol-1 h-1
3.12
0.038
0.15
0.009
a
Excluding the secondary P release by multiplying rPP_PO4,Sec,An by the duration of the anaerobic phase (0.063 P-mmol L-1 h-1 · 2.25 h = 0.142 P-mmol).
Estimated by dividing the maximum volumetric conversion rates by the active biomass concentration at the beginning of the test of 81.4 C-mmol.
c
Estimated: the aerobic consumption of NH4 divided by the duration of the aerobic phase.
d
Estimated: the volumetric aerobic NH4 consumption rate divided by the N-content of the biomass (0.20 N-mol) and the initial active biomass concentration (81.4 C-mmol L-1).
b
As observed in Table 2.2.12, the anaerobic
stoichiometry, in particular the net P-released/Ac uptake
(YAc_PO4,An, PO4/Ac or P/C) ratio of 0.57 P-mol C-mol-1 in
combination with the glycogen consumption to Ac
uptake ratio of 0.57 C-mol C-mol-1 indicates that the
observed biomass activity (physiology) corresponds to
that dominated by PAO metabolism (Table 2.2.3). This
can be confirmed by the relatively low PHV
production/PHB production (YPHV/PHB,An) ratio of 0.12
since PHV/PHB ratios close to or lower than 0.10 are
commonly observed in PAO-enriched systems (Smolders
et al., 1994a) because of the lower glycogen consumption
for Ac uptake when compared to GAO-dominated
systems (Zeng et al., 2003a). Furthermore, when
comparing the observed PO4/Ac ratio with the values
reported in Table 2.2.3 (for enriched EBPR cultures) and
50
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Figure 2.2.5, it appears that PAOs (or their activity) are
dominant in the sludge. However, this is a mere rough
estimation and should be cross-checked with the use of
molecular techniques (chapters 7 and 8). The P/TSS ratio
of the sludge can be also used to assess the expected
PO4/Ac ratio (Figure 2.2.7). Nevertheless, based on the
previous data, the sludge sample shows a satisfactory
EBPR activity.
HAc
PO4-P
PHV
NH4-N
Aerobic
5.0
A)
4.5
16
4.0
14
3.5
12
3.0
10
2.5
8
2.0
6
1.5
4
1.0
2
0.5
0
0.0
PO4-P, NH4-N (mg L-1)
HAc, PHB, PHV, Glycogen (C-mmol L-1)
20
18
PHB
GLY
Anaerobic
0.0
0.5
1.0
1.5
2.0
2.5
3.0 3.5
4.0 4.5
Time (h)
HAc
20
Aerobic
5.0
B)
4.5
HAc (C-mmol L-1)
16
4.0
y=0.0626x
R2=0.9296
14
3.5
12
3.0
10
2.5
8
2.0
y = 4.42 x
R2 = 0.994
y = -3.12 x
R2 =0.996
6
4
0
0.0
1.5
1.0
y = -6.15 x
R2 = 0.997
2
PO4-P (mg L-1)
18
PO4-P
Anaerobic
0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0 3.5
An important tool to assess the data consistency and
quality is the COD balance. In this example, a COD
balance performed to assess the anaerobic EBPR
conversions shows that about 5.83 C-mmol Ac and 2.98
C-mmol of glycogen are consumed, while 7.80 C-mmol
PHAs are produced. Using the COD conversion factor of
32 mg COD C-mmol-1 for both Ac and glycogen and that
of 36 mg COD mg C-mmol-1 for PHAs, a COD balance
error (ΔCOD(%)) of about 0.2 % is estimated using Eq.
2.2.2. A similar COD balance can be performed with the
aerobic COD conversions if a respirometry test is
performed as indicated in Chapter 3.
4.0 4.5
Time (h)
Figure 2.2.8 Graphic representation of the data obtained in the example
of the experimental implementation plan for the execution of a batch
activity test (Test EBPR.AER.2, Table 2.2.10) performed with a lab-enriched
EBPR sludge at 10 °C using synthetic influent at pH 7.0: A) profiles of the
experimental data of interest; and, B) experimental data of Ac and PO4
showing the main trend lines of the conversion rates for the further
estimation of the maximum kinetic rates.
Regarding the anaerobic kinetic rates, the qAc,An and
mPP_PO4,An values of 0.075 C-mol C-mol-1 h-1 and 7.69 ×
10-4 P-mol C-mol-1 h-1, respectively, appear to be lower
than the typical values reported for PAO-enriched
systems of around 0.20 C-mol C-mol-1 h-1 and 2.1 × 10-3
P-mol C-mol-1 h-1, correspondingly (Table 2.2.12).
However, one should realize that the batch activity test
was conducted at a temperature of 10 °C, whereas
previous values reported in literature have mostly been
obtained from tests performed at 20 °C. Thus, for
comparison purposes with tests executed at 20 °C (and
also at other temperatures), the temperature Arrhenius
coefficients (θ) proposed by Meijer (2004) can be applied
(of 1.094 for carbon uptake and 1.071 for anaerobic
maintenance) to estimate the equivalent kinetic rates at
20 °C based on the tests performed at 10 °C. Therefore,
the equivalent kinetic rates at 20 °C are: 0.18 C-mol Cmol-1 h-1 for the maximum acetate uptake rate (estimated
as qAc,20,An = qAc,10,An / θ(10 - 20)) and 1.53 × 10-3 P-mol Cmol-1 h-1 (mPP_PO4,20,An = mPP_PO4,10,An / θ(10 - 20)).
Similarly, for the maximum aerobic specific uptake
rate (qPO4_PP,Ox) of 0.038 P-mol C-mol-1 h-1 observed at 10
°C, an equivalent maximum specific rate of 0.050 P-mol
C-mol-1 h-1 is estimated at 20 °C which is within previous
reported values in literature (Table 2.2.6). As previously
mentioned, biomass growth cannot be directly
determined from the increase in MLVSS since the
potentially low biomass increase will probably fall into
the standard error of the MLVSS analytical
determination technique. Instead, it is estimated based on
the NH4 consumption observed in the test (ensuring that
NH4 is not removed by any other biological or chemical
process). Thus, a maximum specific biomass growth rate
(qBio,Ox) of 0.009 C-mol C-mol-1 h-1 (C-mol new biomass
produced per C-mol of existing biomass) is determined.
After re-calculating the approximate biomass growth rate
from 10 °C at 20 °C (using the Arrhenius temperature
coefficients for biomass growth of 1.081), the estimated
biomass growth rate is 0.020 C-mol C-mol-1 h-1, which is
ACTIVATED SLUDGE ACTIVITY TESTS
in the range of previously reported values for EBPR
sludge (0.016 C-mol C-mol-1 h-1) (Smolders, 1995).
Overall, the stoichiometric and kinetic parameters
obtained in the test are comparable to similar values
previously reported in literature for EBPR cultures
(Table 2.2.3), strongly indicating that the EBPR activity
observed in the test was typical of an enriched EBPR
culture.
2.2.8 Additional considerations
2.2.8.1 GAO occurrence in EBPR systems
Due to the harmful effects that they can have on EBPR
performance, the occurrence of GAOs has been the
subject of extensive research in past years. So far, (i) the
availability of a sole COD source in the influent (either
acetate or propionate), (ii) a temperature higher than 20
°C, and (iii) pH values lower than 7.0 have been
suggested as key factors favouring the presence of GAOs
in (lab-scale) EBPR systems (Filipe et al., 2001b;
Oehmen et al., 2004; Lopez-Vazquez et al., 2009b). The
most characteristic aspects that suggest the presence of
GAOs are anaerobic P-released/C-uptake ratios much
lower than 0.50 P-mol C-mol-1 and incomplete aerobic
PO4 uptake. Despite the apparently frequent occurrence
of GAOs in lab-scale EBPR systems (possibly as a
consequence of operating the lab-scale systems at the
boundaries of the aforementioned parameters: with a
single carbon source, usually acetate, at pH 7.0 and 20
°C), abundant GAO populations have rarely been found
in full-scale municipal EBPR systems (Thomas et al.,
2003; Saunders et al., 2003; Lopez-Vazquez et al.,
2008a; Lopez-Vazquez, 2009; Kong et al., 2006), unless
certain particular conditions (such as the discharge of
industrial effluents) take place (Burow et al., 2007).
Although a low anaerobic P-released/C-uptake ratio
could suggest a higher activity or involvement of GAOs,
as described in the text below, recent observations have
provided evidence that PAOs, under limiting intracellular
poly-P conditions, may be able to perform a similar
metabolism like GAOs under anaerobic conditions but
still achieve aerobic full P-removal (Schuler and Jenkins,
2003; Zhou et al., 2008; Acevedo et al., 2012; Welles et
al., 2015b). Therefore, an anaerobic P-released/C-uptake
ratio considerably lower than 0.50 P-mol C-mol-1 (e.g.
around 0.35 P-mol C-mol-1) observed in EBPR batch
activity tests does not strictly imply that GAOs are more
abundant than PAOs. While the actual mechanisms
influencing the use of a GAO-type metabolism by PAOs
will probably continue to be a matter of future research
efforts, the use of microscopic and molecular techniques
51
(chapters 7 and 8) is strongly recommended to identify
the dominant EBPR microbial population.
2.2.8.2 The effect of carbon source
It is well known that RBCOD fed to the anaerobic stage
(containing mainly volatile fatty acids such as acetate and
propionate) enhances the growth of EBPR biomass
(Comeau et al., 1986; Mino et al., 1998; Oehmen et al.,
2004). Other RBCOD sources, such as glucose, are not
suitable since they appear to enhance the growth of
GAOs or G-bacteria (Cech and Hartman, 1993). Also, it
is assumed that more complex substrates need to be
hydrolysed and fermented to VFA to become available to
EBPR biomass (Wentzel et al., 2008). In this regard, it is
likely that more complex COD sources will not be fully
consumed in the anaerobic stage and will ‘leak’ into the
anoxic or aerobic stages, influencing the EBPR anoxic
and aerobic results (for instance, when executing
combined anaerobic-anoxic-aerobic tests, such as Tests
EBPR.ANOX.2 or EBPR.AER.2). Ideally, for a
satisfactory interpretation of the experimental data
obtained, no COD source may be present in the anoxic or
aerobic stage of an EBPR batch activity test (unless it is
also a subject of interest). In addition, recent
developments hypothesize that, besides the apparently
known PAOs (Candidatus Accumulibacter phosphatis),
other organisms, such as Actinobacteria or S-PAOs
(Kong et al., 2005; Wu et al., 2014), are able to perform
an excessive P-uptake like Accumulibacter using
alternative electron donors (e.g. aminoacids or H2S). For
day-to-day or regular tests on well-known EBPR
systems, the use of synthetic media containing VFA can
be good enough to provide a satisfactory assessment of
EBPR activity, as presented in this chapter. However, the
use and application of more complex COD sources,
which may be undoubtedly present in raw or settled
municipal wastewater, can lead to either sub-optimal
EBPR activity or (yet to be known) different EBPR
metabolisms. The latter has been and will continue to be
a matter of extensive research. Nevertheless, the reader
should be aware that such conditions can lead to results
that may differ from those presented in this chapter.
2.2.8.3 The effect of temperature
While extensive research has been advocated to assess
the temperature dependencies of EBPR cultures,
suggesting that temperatures lower than 20 °C enhance
the growth of PAOs whereas higher temperatures favour
the development of GAOs (Brdjanovic et al., 1997,
1998b; Lopez-Vazquez et al., 2009a), certain
52
observations have indicated that stable EBPR systems
can operate at temperatures higher than 25 °C (Cao et al.,
2009). Though case-specific combinations between
wastewater composition, operating and environmental
conditions will play a major role, the long-term operation
and acclimatization of EBPR cultures to those particular
conditions can also lead to the development and
enrichment of PAO cultures (or similar organisms
sharing the PAOs’ phenotype) capable of performing
stable EBPR at a higher temperature. In this regard, the
identification of these organisms is of major importance
where the methods and techniques presented in Chapter
7 and particularly in Chapter 8 will be needed and
applicable to elucidate the identity of these organisms.
2.2.8.4 The effect of pH
As discussed, pH has a direct influence on EBPR cultures
(Smolders et al., 1994a, Filipe et al., 2001a). During
EBPR batch activity tests, it should be carefully
monitored and well-controlled to obtain reliable data
(avoiding pH fluctuations higher than ± 0.1-0.2).
However, higher pH levels (particularly above pH 8.0)
combined with the presence of RBCOD can lead to
(expectedly) higher anaerobic P release and favour
chemically-induced calcium phosphate precipitation or
even struvite formation depending on the wastewater
composition (NH4MgPO4 or KMgPO4) (Lin et al., 2012;
Mañas et al., 2011). Also, it cannot be disregarded that
the presence of aluminum salts in the wastewater
produced by the discharge of drinking water sludge can
play a role and lead to the precipitation of phosphorus,
particularly if there is enough retention time in the sewer
network. These processes will reduce the biological
availability of phosphorus for PAOs, resulting in the
potential deterioration of the EBPR process as PAOs will
not be able to replenish their intracellular poly-P pools.
Such conditions will lead to different values to those
presented in this chapter. On the other hand, although the
previous process may not be desirable and, thus, should
be avoided in continuous conventional alternating
anaerobic-anoxic-aerobic EBPR systems, it can offer
interesting alternative options for P recovery that may be
worth exploring in view of the essential role of
phosphorus in the food chain and the potential depletion
of the conventional world’s phosphorus sources.
2.2.8.5 Denitrification by EBPR cultures
Denitrifying EBPR sludge has been the subject of
extensive debate since the 1990s. While satisfactory
simultaneous NO3 removal and PO4 uptake has been
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
observed in the anoxic stage of several EBPR systems
(Vlekke et al., 1988; Kuba et al., 1993, 1996, 1997a,
1997b; Wachtmeister et al., 1997; Brdjanovic et al.,
2000; Zeng et al., 2003b), limited or inconsistent
simultaneous denitrification and PO4 uptake has been
observed in other lab- and full-scale studies (KerrnJespersen and Henze, 1993; Hu et al., 2003; Carvalho et
al., 2007; Lopez-Vazquez et al., 2008a). Actually, the
discovery of the existence of two different
Accumulibacter clades (known PAOs) seemed to have
been encouraged by assessing the denitrifying
capabilities of EBPR biomass (Flowers et al., 2008).
Flowers et al. (2008) proposed that the so-called
Accumulibacter clade Type I has a full-denitrifying
capability (able to denitrify from nitrate to di-nitrogen
gas), whereas Accumulibacter clade Type II appears to
be only able to denitrify from nitrite onwards (in line with
the first observations drawn by Kerrn-Jespersen and
Henze, 1993). As suggested by Kuba et al. (1996) and
Lopez-Vazquez et al. (2008a), the simultaneous NO3
removal and anoxic PO4 uptake capability of an EBPR
sludge may be a reflection of both the induction of the
required denitrifying enzymes (nitrate and nitrite
reductase) and the development of EBPR and side
populations able to denitrify. Therefore, when executing
anoxic EBPR activity tests, the relative anoxic EBPR
activities observed can vary widely from practically zero
to considerably high anoxic activities (Table 2.2.8, Table
2.2.9) as a function of the previously discussed exposure
to anoxic conditions and the development of the
denitrifying populations in the lab- or full-scale systems.
2.2.8.6 Excess/shortage of intracellular compounds
Although the depletion of the intracellular poly-P pools
was initially assumed to be a limiting factor for anaerobic
VFA uptake by PAOs (Brdjanovic et al., 1997), later
developments showed that PAOs can utilize higher
amounts of intracellularly stored glycogen as an energy
source for anaerobic VFA uptake to compensate for the
limiting poly-P availability, and still perform satisfactory
fully aerobic P-uptake (Schuler and Jenkins, 2003; Zhou
et al., 2008, Acevedo et al., 2012; Welles et al., 2014,
2015b, 2016, submitted). This is reflected in anaerobic Preleased/C-uptake ratios much lower than 0.50 P-mol
C-mol-1 (Figure 2.2.7) and higher anaerobic glycogen/C
ratios (Table 2.2.3). Based on observations from Schuler
and Jenkins (2003) and Welles et al. (2015b), the shift
from the metabolic utilization of poly-P to glycogen
appears to take place as soon as the biomass total P/TSS
ratio drops below 0.08 mg P g TSS-1 (Acevedo et al.,
2014). Furthermore, Welles et al. (2015b) have observed
ACTIVATED SLUDGE ACTIVITY TESTS
that the maximum kinetic rates of the two known
Accumulibacter clades (Flowers et al., 2008) appear to be
affected in a different manner depending on the
intracellular poly-P availability, with Accumulibacter
Type I more affected than Accumulibacter Clade Type II.
This can explain potential deviations from those
presented in this chapter, which can be supported by the
determination of the intracellular poly-P and glycogen
contents or at least by the estimation of the biomass
P/TSS ratio.
2.2.8.7 Excessive aeration
The exposure of EBPR biomass to extended aeration
periods (e.g. longer than 12-24 h) can lead to the
sequential utilization of PHAs, glycogen and intracellular
poly-P under aerobic conditions (Brdjanovic et al.,
1998c; Lopez et al., 2006), as a consequence of the
biomass needs to cover their aerobic maintenance
requirements. Consequently, it will lead to an eventual
aerobic P release as soon as the intracellularly stored
poly-P pools start to be hydrolysed. Thus, if EBPR sludge
samples are aerated for extensive periods of time prior to
the execution of the batch activity tests (e.g. overnight or
during transportation), it is likely that a lower EBPR
activity will be observed due to the potentially low(er)
poly-P and glycogen contents of the sludge. Therefore,
over-aeration
conditions
should
be
avoided.
Furthermore, it deserves particular attention for the
satisfactory operation of full-scale activated sludge
EBPR systems because excessive aeration periods during
low loading conditions (e.g. weekends or holidays
periods) can result in undesired aerobic P release,
affecting the treated effluent quality.
2.2.8.8 Shortage of essential ions
Though it may be trivial, the presence of macro- and
micro-nutrients in the right concentration and (bio-)
availability is essential for EBPR activated sludge
systems. For instance, potassium, magnesium, iron, and
calcium, among others, are rather important to regulate
the microbial EBPR metabolism and support the storage
of intracellular compounds (Brdjanovic et al., 1997;
Burow et al., 2007; Barat et al., 2008). Their absence, for
instance of potassium, can lead to the deterioration of the
EBPR process (Brdjanovic et al., 1997), but their excess
can influence the metabolism of PAOs inducing a GAO
metabolism (Jobaggy et al., 2006; Barat et al., 2008),
possibly due to the chemical precipitation of phosphorus
with the aforementioned elements. This will reduce their
bio-availability and therefore the aerobic replenishment
53
of poly-P pools. The absence or presence in excessive
concentrations of the aforementioned elements should be
checked if one suspects that their concentrations differ
from those regularly observed in municipal wastewater
treatment systems.
2.2.8.9 Toxicity/inhibition
A limited number of compounds have been identified to
be toxic or inhibiting to the EBPR process. The presence
of nitrate or nitrite in the anaerobic zone is considered
harmful for the EBPR process since they enhance the
activity of ordinary denitrifying organisms that can
consume the available RBCOD to the detriment of PAOs
(Wentzel et al., 2008). Also, the presence of nitrite in the
aerobic zone in concentrations as low as 6-8 mg L-1 has
been proven to inhibit PAOs (Saito et al., 2004), which
can worsen the negative effects at lower pH due to the
increase in free nitrous acid (FNA) (Zhou et al., 2008;
Pijuan et al., 2010). The latter favours the occurrence of
GAOs over PAOs since GAOs appear to be more tolerant
to the presence of FNA (Pijuan et al., 2011), resulting in
the deterioration of the EBPR process. Salinity is another
factor that may inhibit the activity of PAOs in the shortterm (hours) (Welles et al., 2014, 2015a). Welles et al.
(2014) observed that, after a short-term exposure (hours),
chloride concentrations as low as 10,000 mg NaCl L-1 (1
% salinity) can lead to more than 50 % inhibition of the
anaerobic metabolism of PAOs and practically fully
inhibit the aerobic metabolism. Similar circumstances
can take place due to a sudden saline intrusion into the
sewage or saline intrusion to the plant (particularly in
coastal regions), or caused by industrial effluent
discharges. Nevertheless, the long-term exposure of
EBPR sludge to high salinity concentrations (> 35,000
mg NaCl L-1, 3.5 % salinity) can enhance the
acclimatization of EBPR biomass which can become
salt-tolerant and be able to perform satisfactory EBPR at
salinity concentrations equivalent to those observed in
seawater. Another compound that is potentially
inhibitory or toxic to EBPR sludge is H2S, which may be
formed in the sewage (due to saline intrusion or industrial
discharges) or in the anaerobic stage of the EBPR
activated sludge system. Its presence can be rather
inhibitory to the anaerobic metabolism of PAOs at
concentrations as low as 20-25 mg H2S L-1 leading to 50
% inhibition (Saad et al., 2013; Rubio-Rincon et al.,
2016, submitted). Overall, inhibiting effects will be
reflected in limiting or sub-optimal EBPR activity. If
feasible, the sludge can be washed in a mineral solution
(as explained in Section 2.2.3.5) to avoid the potentially
inhibiting or toxic compounds.
54
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.3 BIOLOGICAL SULPHATE-REDUCTION
2.3.1 Process description
Sulphate (SO42-) is naturally present in surface water and
ground water and depending on the geographical
location, its concentrations in drinking water supply
networks can vary. Consequently, the presence of other
sulphur compounds namely organic sulphides, including
mercaptans, dimethyl sulphides and dimethyl disulphides
is also very common in domestic wastewater. The
concentration of sulphate in domestic sewage usually
ranges from 20 to 60 mg L-1 (Moussa et al., 2006).
However, the sulphate concentration in sewage can reach
as high as 500 mg L-1, due to the discharge of sulphaterich industrial effluents, seawater-based toilet flushing or
intrusion of saline water into the sewer (Lens et al., 1998;
Chen et al., 2010; Ekama et al., 2010).
The metabolism of sulphate-reducing bacteria (SRB),
beside its applications in domestic sewage treatment, can
be exploited beneficially in specific industrial effluent
treatment processes that generate sulphate-rich
wastewater (Lens et al., 1998; Muyzer and Stams, 2008).
Such industries are, among others, potato starch
production, pulp and paper mills, food and fermentation
industries and seafood processing facilities. In
wastewater treatment applications, sulphate is usually
completely reduced to sulphide, as this conversion yields
the highest free Gibb’s energy (ΔG°΄, Table 2.3.1).
Table 2.3.1 Sulphate transformation reactions performed by SRB
(Jørgensen, 2006; Liamleam and Annachhatre, 2007).
ΔG°΄ (KJ mol-1)
Reactions
SO + 4H2 + H → HS + 4H2O
24
+
-
-152.2
SO + H2 + 2H → HSO + H2O
+19.7
HSO3- + 3H2 → HS- + 3H2O
-171.7
24
+
3
3HSO3- + H2 + H+ → S3O62- + 3H2O
-46.3
S3O62- + H2 → S2O32- + HSO3- + H+
-123.0
S2O32- + H2 → HS- + HSO3SO42- + 2H+ + ATP → APS + PPi
-2.1
+46.0
PPi + H2O → 2Pi
-21.9
APS + H2 → HSO3- + AMP + H+
-68.0
Dissimilatory sulphate-reduction is the most
important anaerobic process in many different
environments (Balk et al., 2008). The initial step of
biological sulphate-reduction involves the transfer of
exogenous sulphate through the bacterial cell membrane
into the cell. The sulphate dissimilation process proceeds
via the action of adenosine triphosphate (ATP)
sulphurylase (Figure 2.3.1). ATP produces the highly
activated molecule adenosine phosphosulphate (APS),
and pyrophosphate (PPi) in the presence of sulphate,
which yields inorganic phosphate. Further, APS is
rapidly converted to bisulphite (HSO3-) by the
cytoplasmic enzyme APS reductase. Pyrophosphate is
hydrolyzed and the sulphate moiety of APS is reduced to
bisulphite, together with adenosine monophosphate
(AMP). Bisulphite in turn may be reduced via a number
of intermediates to form the sulphide ion. Bisulphite is
reduced to bisulfide (HS-) via bisulphite reductase. By
another mechanism, bisulphite reduction via the enzymes
bisulphite reductase, trithionate reductase and thiosulfate
reductase yields trithionate (S3O62-) and thiosulfate
(S2O32-) as free intermediates. The physiology and growth
of these bacteria has been studied in depth and is well
documented (Cypionka, 1987; Gibson, 1990; Hansen,
1994; Rabus et al., 2006).
The organisms responsible for the reduction of
sulphur compounds belong to both bacteria and
prokaryotes (Postgate, 1965; Muyzer and Stams, 2008);
nonetheless, in literature, the term SRB is used. In this
chapter, the term SRB includes both bacteria and
prokaryotes. The bacterial sulphate reducers are
categorized
into
different
branches,
the
Deltaproteobacteria with more than 25 genera, the
Gram-positive bacteria that include Desulfotomaculum
and Thermodesulfobium and Gram-negative sulphate
reducers that include Thermodesulfobacterium and
Thermodesulfatator (Mori et al., 2003; Moussard et al.,
2004; Balk et al., 2008). In general, SRB are found
present and active in sewerage systems and wastewater
treatment plants. Several authors have also reported SRB
activities in freshwater, marine, hypersaline and
oil/hydrocarbon polluted sites (Cravo-Laureau et al.,
2004; Almeida et al., 2006; Kjeldsen et al., 2007).
SRB are facultative anaerobes that live in oxygenfree or depleted environments and utilize sulphate as a
terminal electron acceptor to produce hydrogen sulphide
(H2S) as one of its metabolic end products. SRB can
survive under extreme environmental and operating
conditions over a rather wide range of pH (4.0 to 9.5),
temperature (25-75 °C) and pressures of up to 500 atm
(Madigan et al., 2009; Tang et al., 2009). Sulphate is
redox sensitive and the production of sulphide is an
indicator of SRB activity which depends on several
factors including sulphate concentrations, the
ACTIVATED SLUDGE ACTIVITY TESTS
55
concentration of organic matter/nutrients, pH and
temperature, among others. In sewerage and municipal
wastewater treatment plants, the presence of SRB is
considered to be undesirable, due to corrosion and
instability in methanogenic activity, causing an
insufficient digestion process (Oude Elfrenink et al.,
1994). SRB have been recognized as the major
microbiologically-influenced corrosion-causing bacteria
in oil/gas pipelines and sewer systems (Al Abbas et al.,
2013). Microbiologically-influenced corrosion is
aggravated due to the synergistic interaction of different
microbes such as iron and manganese-reducing bacteria,
carbon dioxide-reducing bacteria that co-exist and
through cooperative metabolism with SRB (Little and
Lee, 2007). According to Kjeldsen et al. (2004), the
presence of SRB in activated sludge is of interest because
sulphate-reduction can have negative effects on the
wastewater treatment plant operation. Other negative
side-effects of the activity of SRB is that the produced
sulphide will inhibit other main microorganisms
involved in the treatment process, such as methanogenic
bacteria (MET), phosphorus-accumulating organisms
(PAO), nitrifiers and others. Exceedingly high sulphide
levels can be toxic for microorganisms performing
methanogenesis and sulphate-reduction. High sulphide
concentration can also have a deteriorating effect on the
activated sludge floc structure, for instance,
deflocculation, by reducing Fe(III) to Fe(II) as FeS
(Caccavo et al., 1996; Nielsen and Keiding, 1998). The
authors attributed this phenomenon to the better
flocculating properties of Fe(III) rather than Fe(II),
mainly due to its valance and lower solubility.
On the other hand, sulphide production can lead to
the growth of filamentous bacteria and result in
concomitant sludge bulking (Yamamoto et al., 1991;
Zeitz et al., 1995; Kjeldsen et al., 2004).
Assimilatory pathway
Dissimilatory pathway
Sulphate
Sulphate
ATP
ATP
ATP sulphurylase (Sat)
2Pi
PPi
Sat
2Pi
Adenosine 5’phosphosulphate (APS)
PPi
Adenosine 5’phosphosulphate (APS)
ATP
ADP
Adenylyl sulphate
kinase (CysC)
2eAPS reductase (AprAB)
3’- Phosphoadenosine 5’phosphosulphate (PAPS)
AMP
ThioredoxinRed
Sulphite
PAPS reductase (CysH)
ThioredoxinOx
6eSulphite
NADPH or FdRed
NADP+ or FdOx
Assimilatory
sulphite reductase
Sulphide
O-acetylserine
sulfhydrylase
Cysteine
Figure 2.3.1 Prokaryotic assimilatory and dissimilatory pathways for sulphate-reduction (Grein et al., 2013).
Dissimilatory sulphite
reductase DsrAB/DsrC
Sulphide
56
100
(A)
HS90
H2S
S2-
80
Percentage (%)
70
60
50
40
30
20
10
0
5
4
6
7
9
pH
8
11
10
1.2
12
13
14
(B)
1.1
1.0
0.9
0.8
0.7
HSO4-
0.6
0.5
0.4
0.3
Eh (V)
Driven by the potable water supply stress mitigated
by using seawater for toilet flushing, the research group
of Hong Kong University of Science and Technology
(HKUST) recognized the benefits of sulphate present in
saline sewage and successfully developed a process that
takes a big advantage of sulphate and SRB: the SANI®
process (Sulphate-reduction, Autotrophic denitrification
and Nitrification Integrated process); the only technology
that applies SRB for municipal wastewater treatment
purposes. Its application has led to lower sludge
production, higher coliform (pathogen) and heavy metal
removal while using less space and energy (Wang et al.,
2009; Abdeen et al., 2010). In addition, SRB-based
processes have been used as a pre-treatment step to
enhance the digestion process. Recently, Daigger et al.
(2015) tested a pilot-scale membrane bioreactor (MBR)
to remove elemental sulphur from an anaerobically pretreated pulp and paper effluent containing high
concentrations of dissolved sulphide. Although SRB play
a major role in treatment plants, they are not frequently
studied in domestic wastewater treatment. To understand
the positive and negative effects of SRB activity in a
domestic wastewater treatment plant, activity tests
should be conducted. Batch tests can be useful to estimate
to what extent SRB are present and active in sewage and
treatment plants and to provide an insight to develop
measures that can stimulate or suppress SRB activity,
depending on the chosen process design and operational
conditions. Therefore, from a process performance
perspective, the stability and continuation of the
treatment process is highly dependent on (among others)
the concentration of the sulphide (and pH) present in the
liquid phase.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
SO42-
0.2
0.1
0.0
H2S(aq)
-0.1
-0.2
-0.3
-0.4
HS-
-0.5
2.3.2 Sulphide speciation
-0.6
Sulphide can be present in wastewater in various
states (H2S, HS- and S2-). The unionized H2S is known to
have the strongest inhibitory effect, due to its ability to
permeate the cell membrane. pH is the main factor that
determines the proportion of S2- present in wastewater.
Figure 2.3.2A indicates the relation between pH and
sulphide speciation. As the pH of wastewater is usually
around 7.6, sulphide will mostly be present as HS-. The
Eh-pH diagram depicted in Figure 2.3.2B shows the
dominant aqueous species and stable solid phases on a
plot defined by the Eh and pH axes. It is noteworthy that,
under anaerobic conditions, the zero valent dissolved and
suspended sulphur may exist in aqueous solutions as
colloidal sulphur or as metal-ligated or free
polysulphides and hydropolysulphides (Kamyshny et al.,
2008).
-0.8
-0.7
S2-
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14
pH
Figure 2.3.2 Sulphide speciation in wastewater: (A) pH influence at 25 °C,
as adapted from Rintala and Puhakka (1994); and (B) Eh-pH diagram of
various S species in aqueous solution (FACT database, adapted from Bale et
al., 2002).
Sulphide speciation can also be influenced by the
partition coefficient value, i.e. the ratio of concentration
between the gas phase and the liquid phase, which is also
heavily pH and temperature-dependent. Other
parameters which might influence sulphide speciation are
temperature and salt concentration.
ACTIVATED SLUDGE ACTIVITY TESTS
57
The distribution of sulphide in the gas phase (g) and
liquid phase (l) can be represented by the following
equation (Hulshoff Pol et al., 1998):
SH2S = α · CH2S
is usually added. Liu and Peck (1981) determined
appropriate electron donors for SRB when grown as a
pure culture. Several electron donors have been studied
as energy and carbon sources for SRB (Table 2.3.2).
Eq. 2.3.1
Where, α (alpha) is the dimensionless distribution
coefficient for H2S liquid-gas phase equilibrium. In the
liquid phase, SH2S, H2S exists as uninonized H2S or in its
ionized forms (as bisulfide, HS- or sulphide, S2-). Even
small variations in the pH value will significantly affect
the free H2S concentrations. Again, the Henry's law
constant for H2S at 25 °C is ~3.4 mg L-1 atm-1, which
indicates the large volatility of this species. The
dependence of α on the temperature and pKa values (pKa
= - log Ka) is shown in Figure 2.3.3.
Table 2.3.2 Compounds used as energy substrates by SRB (adapted from
Hansen, 1993).
Compound
Substrate
Inorganic
Monocarboxylic acids
Hydrogen, carbon monoxide, etc.
Lactate, acetate, butyrate, formate, propionate,
isobutyrate, 2- and 3- methylbutyrate, higher
fatty acids up to C18, pyruvate.
Succinate, fumarate, malate, oxalate, maleinate,
glutarate, pimelate.
Methanol, ethanol, propanol, butanol, ethylene
glycol, 1,2- and 1,3-propanediol, glycerol.
Glycine, serine, cysteine, threonine, valine,
leucine, isoleucine, aspartate, glutamate,
phenylalanine.
Choline, furfural, oxamate, fructose, benzoate,
2-, 3- and 4-OH-benzoate,
cyclohexanecarboxylate, hippurate, nicotinic
acid, indole, anthranilate, quinoline, phenol,
p-cresol, catechol, resorcinol, hydroquinone,
protocatechuate, phloroglucinol, pyrogallol, 4OH-phenyl-acetate, 3-phenylpropionate, 2aminobenzoate, dihydroxyacetone
Dicarboxylic acids
Alcohols
5
20
Amino acids
4
15
10
2
Miscellaneous
8
KH S (10 )
2
Alpha
3
5
1
0
0
0
10
20
30
40
50
Temperature (˚C)
Figure 2.3.3 Temperature dependency on the distribution coefficient
alpha (●) and dissociation constant pKa (●) of H2S (adapted from
Hulshoff Pol et al., 1998).
2.3.3 Effects of environmental and
operating conditions on SRB
2.3.3.1 Carbon source
To obtain energy for growth and maintenance, SRB can
oxidize a wide range of substrates acting as the electron
donor that include, among others, hydrogen, alcohols,
fatty acids, aromatic and aliphatic compounds (Liamleam
and Annachhatre, 2007), while sulphate is used as an
external electron acceptor. In certain (industrial) plants,
when the wastewater contains an insufficient electron
donor/carbon source and there is interest in promoting the
removal of organics by SRB, an external carbon source
Most electron donors are products of fermentation,
monomers or cell components from other sources. The
three major criteria for selecting a suitable electron donor
for sulphate-reduction process are: (i) efficiency of
sulphate removal in the effluent complemented by a low
COD, (ii) the electron donor availability, and (iii) cost of
unit sulphate converted to sulphide (van Houten et al.,
1996). With greater interest in the development of
sustainable SRB technologies, hydrogen has been
identified as one of the most important substrates for
SRB. Desulfovibrio species have a high affinity for
hydrogen, and this is considered to be the reason why
they are able to out-compete hydrogenotrophic
methanogens in sulphate-rich environments (Widdel,
2006).
Recently used energy sources for biological sulphatereduction for industrial applications include complex
organic carbon sources. Van Houten et al. (1996) used
synthetic gas (mixtures of H2, CO, and CO2) as the energy
source in laboratory scale gas-lift reactors and showed
58
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
biological sulphate-reduction by SRB. Numerous
organic waste matrices (sources) have also been used as
carbon sources and electron donors. These include leaf
mulch, wood chips, sewage sludge, sawdust, compost,
animal manure, whey, vegetable compost, and other
agricultural wastes (Liamleam and Annachhatre, 2007).
Lactate and molasses, though cost-effective, are not
completely oxidized by SRB, generating high COD in the
effluent. Hydrogen and ethanol, despite being more
expensive, are still being used for sulphate loads greater
than 200 kg SO42- h-1. However, due to safety reasons,
ethanol is preferred to hydrogen. Interestingly, one of the
first carbon sources to be considered was waste sewage
sludge (Butlin et al., 1956) and COD is usually present
in relatively higher concentrations in municipal
wastewater, suggesting that for municipal wastewater
treatment applications no external electron donors may
be needed. During the fermentation of COD in the
sewerage system or in the anaerobic zones of wastewater
treatment plants, volatile fatty acids (VFA) such as
acetate and propionate may contribute to most of the
readily biodegradable COD (RBCOD). Other
microorganisms like methanogens or acetogens can also
utilize VFA as a carbon source under anaerobic
conditions, competing against SRB for VFA. The
interactions between SRB, methanogens and acetogens
depend on the type of VFA and other substrates present
in the wastewater (Figure 2.3.4).
From the equations and Gibb's free energy values
(ΔG°΄) shown in Table 2.3.3, it can be observed that SRB
can utilize a broad range of carbon sources (lactate,
hydrogen, acetate, propionate and butyrate), whereas
methanogens rely mostly on hydrogen and acetate, while
acetogens can utilize lactate, propionate and butyrate.
Thus, SRB will compete against acetogens for lactate,
propionate and butyrate, and against methanogens for
hydrogen and acetate. Overall, SRB can have bigger
advantages over other anaerobic organisms, as long as
there is sufficient sulphate available.
Table 2.3.3 Typical reactions for sulphur-reducing bacteria, methanogens
and acetogens: ΔG°΄ (kJ mol-1) values are adapted from Thauer et al.
(2007).
ΔG°΄(kJ mol-1)
Reaction
Sulphate-reducing bacteria
Lactate- + 0.5SO42- → Acetate- + HCO3- + 0.5HS-
-80.2
4H2 + SO42- + H+ → HS- + 4H2O
-36.4
Acetate + SO4 → 2HCO3 + HS
-47.6
1.33Propionate- + SO42-→1.33Acetate-+ 1.33HCO3- + 0.75HS-+ 1.33H+
-50.3
2Butyrate- + SO42- → 4Acetate- + HS- + H+
-55.6
2-
-
-
-
Methanogens
4H2 + HCO3- + H+ → CH4 + 3H2O
-33.9
Acetate- + H2O → CH4 + HCO3-
-31.0
Acetogens
Lactate- + 2H2O → Acetate- + HCO3- + H+ + 2H2
Propionate + 3H2O → Acetate + HCO3 + H + 3H2
-
-
-
Butyrate- + 3H2O → 2Acetate- + H+ + 2H2
+
-4.2
+71.6
+9.6
2.3.3.2 COD to SO42- ratio
As shown in Table 2.3.3, the oxidation of COD is
coupled to the reduction of sulphate, in a COD/SO42- ratio
of 0.67 g COD g SO4-1 (Khanal, 2008). Thus, a complete
COD oxidation by SRB can only be achieved if the
conditions are not sulphate-limiting (i.e. the COD/SO42ratio is 0.67 or lower). For medium-strength domestic
wastewaters (for e.g. soluble COD concentrations of
approximately 300 mg COD L-1), the minimal amount of
sulphate required in the influent can easily be achieved
due to the intrusion of brackish water into the sewer,
seawater-based toilet flushing and/or discharge of
sulphate-rich (industrial) wastewaters. On the other hand,
a higher COD/SO42- ratio will favour the growth of
methanogenic bacteria that will consume part of the
available RBCOD.
2.3.3.3 Temperature
SRB
Propionate
SRB
Acetate
Acetogens
CO2
Methanogens
Figure 2.3.4 The competition between sulphate-reducing bacteria (SRB),
methanogens and acetogens for VFA in municipal wastewater.
The sulphate-reduction rate of SRB and their doubling
time are highly dependent on temperature (Hulshoff Pol
et al., 1998; Pikuta et al., 2000). The doubling time of
microorganisms capable of performing intense
sulphidogenesis has been observed to drastically vary
between 10 and 118 h within a temperature range of 30
to 60 °C (Pikuta et al., 2000). Table 2.3.4 shows the
optimum temperature for the growth of some SRB.
ACTIVATED SLUDGE ACTIVITY TESTS
59
Temperature (°C)
SRB
Desulfobacter
Desulfobulbus
Desulfomonas
Desulfosarcina
Desulfovibrio
Thermodesulforhabdus norvegicus
Desulfotomaculum luciae
Desulfotomaculum solfataricum
Desulfotomaculum thermobenzoicum
Desulfotomaculum thermocisternum
Desulfotomaculum thermosapovorans
Desulfacinum infernum
Range
Optimum
28–32
28–39
–
33–38
25–35
44–74
50–70
48–65
45–62
41–75
35–60
64
–
–
30
–
–
60
–
60
55
62
50
–
are of practical interest. A typical example showing the
effect of pH and temperature on the doubling time of
SRB is illustrated in Figure 2.3.5.
140
A)
B)
120
100
Doubling time (h)
Table 2.3.4 Temperature range for some SRB (Tang et al., 2009).
80
60
40
20
The effect of temperature on bacterial sulphatereduction rates can be evaluated using the Arrhenius
model (Isaksen and Jørgensen, 1996). The activation
energy for sulphate-reduction can be determined by
plotting the logarithm of the sulphate-reduction rate
versus the inverse of temperature as follows:
ln rSO4.An = ln A +
‒Ea
RT
Eq. 2.3.2
Where, Ea is the activation energy (J mol-1), rSO4,An is
the sulphate-reduction rate (nmol cm-3 d-1), Where, Ea is
the activation energy (J mol-1), k is the sulphate-reduction
rate (nmol cm-3 d-1), A is a constant, R is the molecular
gas constant (8.314 J K-1 mol-1), and T is the absolute
temperature (K).
2.3.3.4 pH
Another parameter that may influence the competition
between SRB and methanogens is pH. pH affects the
different metabolic processes of SRB, methanogens and
acetogens. Also, pH affects the extent to which VFA and
sulphide are inhibitory or toxic to SRB. A slight increase
in pH (e.g. from 7.6 to 7.8) helps SRB to become
dominant. Both VFA and sulphide tend to be more
inhibitory or toxic at low pH values because the
protonated species of VFA and total solids (TS) become
abundant at these pH and, due to their neutral charge,
they can freely diffuse through the membranes of the
organisms and affect their intracellular constituents such
as enzyme mechanisms. In any case, methanogens are
more inhibited by the presence of sulphide than SRB.
Therefore, batch activity tests regarding the VFA
content, COD/SO42- ratio and pH effect on SRB activity
0
8.0 8.2 8.4 8.6 8.8 9.0
30 35 40 45 50 55 60 65
pH
Temperature (˚C)
Figure 2.3.5 Influence of (A) pH and (B) temperature on the doubling time
of bacterial strains capable of performing high rate sulphidogenesis
(adapted from Pikuta et al., 2000).
Pikuta et al. (2000) showed the influence of pH on
the doubling time of a strain (referred to as S1T) capable
of showing high rate sulphidogenesis. As illustrated in
Figure 2.3.5, the optimum pH for growth was between
8.0 and 9.15, while the doubling times varied between 20
and 58 h depending on the initial pH values. The
doubling time of another sulphate-reducing strain HHQ
20
isolated
using
enrichments
with
hydroxyhydroquinone and sediment from Venice was
lowest (40 h) in the pH range of 6.9-7.2, while at pH 5.0
and 8.0 the doubling times were threefold higher
(Reichenbecher and Schink, 1997).
2.3.3.5 Oxygen
SRB are known to be anaerobic organisms. However,
sulphate-reduction activity has been reported in oxic
regions of microbial mats and sea sediments but, to date,
there is no known pure culture capable of performing
dissimilatory sulphate-reduction in the presence of
oxygen concentrations larger than 1 μM (Cypionka,
2000; Kjeldsen et al., 2004). Kjeldsen et al. (2004)
performed short and long term oxygen exposure
experiments and monitored sulphate-reduction rates
under anoxic incubation conditions (Figure 2.3.6).
60
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Time (h)
0
10
20
30
40
under steady-state conditions for at least three SRT.
However, if one would like to envisage the evolution of
the rate over time, after changing the bioreactor’s process
conditions, the rate should be checked more frequently.
For instance, depending on (i) the research question to be
addressed, (ii) the operating conditions, (iii) substrate and
(iv) toxic compounds (if any) present, SRB activity tests
should be performed in triplicates, at least once every two
weeks.
50
60
1.5
1.0
Sulphide (mmol)
50
0.5
40
Sulphide (mmol)
2.0
0.0
30
2.3.4.1 Estimation of volumetric and specific rates
20
10
0
0
20
40
60
80
100
120
140
Time (h)
Figure 2.3.6 Sulphide production profiles in batch incubations of sludge
samples exposed to 0 h (●), 33 h (●), 73 h (●), and 121 h (●) of
constant aeration. The symbol (●) represents a sample that was sterilized
and exposed to 121 h of aeration. The inset represents the initial linear
phases of the respective curves of the same experiment (adapted from
Kjeldsen et al., 2004).
Table 2.3.5 gives an overview of parameters of interest.
The volumetric rates of sulphate-reduction or sulphide
production can be calculated from the highest slope of the
exponential phase for each experiment. The equations for
the volumetric and specific activity rates are:
rSO4,An = ln
SSO4,final
/ Time
SSO4,initial
Eq. 2.3.3
qSO4,An = Volumetric activity / Initial biomass concentration
or,
qSRB,SO4,An =
rSO4,An
XSRB
Eq. 2.3.4
The sulphate-reduction rates, calculated from the
initial linear increase in sulphide concentrations,
decreased from 0.24 to 0.04 μM h-1 when the aeration
time was increased from 3 to 121 h. This was attributed
to temporary inactivation of an increasing fraction of the
SRB population. As shown in Figure 2.3.6, the sulphide
production curves follow a biphasic profile, i.e. an initial
linear increase resulting from the activity of a constant
number of SRB followed by an exponential increase
representing the reactivation of SRB. The authors, who
monitored the oxygen concentration in the bulk liquid
phase and not inside the sludge flocs, raised an important
question on whether the SRB were protected from
oxygen exposure in anoxic micro-niches created by high
respiratory activity and diffusion resistance inside the
flocs during aeration experiments. The results from their
experiment suggest that SRB are capable of adapting well
to the situation and short exposure times to oxygen do not
necessarily affect their sulphate-reduction activities.
Where, SSO4,final is the final concentration of sulphate
(mg SO42- L-1) and SSO4,initial is the initial concentration of
sulphate (mg SO42- L-1) in a batch activity test,
respectively.
2.3.4 Experimental setup
2.3.4.2 The reactor
To estimate the sulphate-reduction rate or sulphide
production rate of a specific activated sludge reactor, one
activity test should be performed if the reactor is operated
To assess the presence and activity of SRB, batch tests
should be conducted under anaerobic conditions either
using airtight reactors (see Sections 2.2.2.1 and 2.2.4.1)
Table 2.3.5 Stoichiometric and kinetic parameters of interest for SRBcontaining activated sludge (AB: active biomass).
Parameter
Stoichiometric
SO42- reduction to VFA
consumption ratio
Kinetic
Specific VFA
consumption rate
Specific SO42uptake rate
Symbol
Typical units
YSO4/VFA,An S-mol C-mol-1
mg SO42- mg VFA-1
qSRB,VFA,An C-mol C-mol-1 h-1
mg VFA mg AB-1 h-1
qSRB,SO4,An S-mol C-mol-1 h-1
mg SO42- mg AB-1 h-1
ACTIVATED SLUDGE ACTIVITY TESTS
or serum bottles (Figure 2.3.7). The reactor(s) or serum
bottles used for the execution of tests must have the
required means to: (i) avoid oxygen intrusion under
anaerobic conditions, (ii) provide satisfactory mixing
conditions, (iii) maintain an adequate and desirable
temperature, (iv) control pH, and (v) have additional ports
for sample collection and addition of influent, solutions,
gases and any other liquid medium or substrate used in
the test.
61
sample. By maintaining multiple vials for the same test
condition, the possibility of contamination during
sampling can be completely avoided and the gas and
liquid phase volumes will remain constant. As explained
previously, sufficient care should be taken to provide the
strictly anaerobic environments required for SRB
growth. Apart from measuring sulphide in the liquid
phase, the gas-phase hydrogen sulphide profiles should
also be monitored in order to close the sulphur balance
within the test bottles.
2.3.4.3 Mixing
When airtight reactors are used, mixing conditions can be
provided as described in Section 2.2.2.1 which covers
EBPR. When serum bottles are used, mixing conditions
are provided by means of an orbital shaker in which
several batch bottles can be placed and operated at the
same time. Proper mixing should be ensured and
stratification, which might result due to variations in
settling capacity of activated sludge, should be avoided.
The shaker speed should be adjusted in such a way that
no disturbance is caused to the floc structure.
2.3.4.4 pH control
Figure 2.3.7 Batch serum bottles containing SRB sludge placed in a shaker
(photo: van den Brand, 2015).
Moreover, several serum bottles can be used in
parallel, as an alternative to determine the kinetic rates in
batch activity tests. This practice is known as using
sacrificial bottles. They are of particular interest when
pure SRB cultures are to be used to avoid contamination
but also to assess the kinetic conversion rates under strict
anaerobic conditions and without jeopardizing the gasliquid equilibrium phase, which can be easily altered
when sampling. In this case, when using sacrificial
bottles, all the batch experiments should be started with
identical initial biomass concentrations that will allow
the estimation of sulphate-reduction rates and sulphide
production rates. In order to perform these tests and to
maintain the purity of the original SRB (when required),
different glass vials should be used maintaining similar
experimental conditions and sacrificing one vial for each
pH affects the different metabolic processes of SRB,
MET and AC and therefore needs to be carefully
controlled and monitored. In reactors, the pH can be
controlled as described in Section 2.2.2.1 on EBPR. In
serum bottles, pH may be controlled by the presence of a
buffering agent in the medium. This can be done by the
addition of buffers such as CaCO3 or phosphate buffer
during the start of the incubation. Some authors have
recommended the addition of sodium bicarbonate buffers
or Tris buffer made from an amine base (HOCH2)3CNH2
to control pH involving microbiological works. In
bioreactors, solutions made of KOH, NaOH, NH4OH,
Ca(OH)2, HCl or H2SO4 can be automatically added to
maintain the pH (Mohan et al., 2005).
2.3.4.5 Temperature control
The sulphate-reduction rates of SRB and their doubling
time are dependent on temperature (Hulshoff Pol et al.,
1998; Pikuta et al., 2000). Therefore, it is recommended
to operate the reactors with automatic temperature
control, following the procedures and equipment
described in Section 2.2.2.1 on EBPR. For experiments
with serum bottles, often a water bath, incubator or a
small room with temperature control is used. This makes
it possible to control the temperature of many serum
bottles simultaneously.
62
2.3.4.6 Sampling and dosing ports
When experiments are conducted in reactors, the
requirements for the sampling and dosing ports will be
the same as described in Section 2.2.2.1 on EBPR. When
serum bottles are used, gas or liquid samples are taken by
syringes with hypodermic needles through the rubber
stoppers. If there is enough headspace pressure in the
flask, no oxygen intrusion into the system will occur.
Alternatively, to avoid the formation of micro-holes in
the stopper or septa due to repeated piercing with the
needle during sampling, a three-way plastic valve fitted
to a needle can be permanently fixed to the septa.
2.3.4.7 Sample collection
For specific cases, if there is interest to measure SRB
activity in the sewage, raw influent samples should be
taken. In this case, most of the SRB activity in the
sewerage will be linked to SRB present in the biofilm
attached to the sewer conduits so that the activity of SRB
measured in the raw influent in the lab may not be
representative of the actual activity taking place in the
sewerage. Specific protocols should be followed while
collecting biofilm sewer samples (Flemming et al.,
2000).
At pilot and full-scale treatment plants, sludge
samples can be collected in different places where the
SRB activity needs to be addressed: primary settling
tanks, anaerobic selectors, anoxic or aerobic tanks,
thickeners or anaerobic digesters. In cases where
experiments are conducted with laboratory-enriched
SRB cultures, it is also possible to convert the continuous
bioreactor or sequencing batch reactor into a batch
operation mode to measure the activities. This is often
considered to be beneficial as the setup is usually
adequately equipped, and often more options for
controlling and sampling are present, as compared to
regular batch flasks. The disadvantage of performing a
batch experiment in the 'parent' reactor (as is generally
applicable for all microbial cultures of interest) is the
possible disturbance of the system because of (a
combination of) several factors such as the withdrawal of
mixed liquor due to additional sampling, the application
of different operating conditions required by the batch
tests, changing the substrate composition and
concentration, or changing the duration of phases.
2.3.4.8 Media
When activated sludge is used as a source of SRB, real
wastewater can be used as substrate, following the
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
procedures described in Sections 2.2.2.4 and 2.2.3.3 on
EBPR. The medium composition depends heavily on the
objective of the test. For determining the rates of SRB
under the conditions applied in the parent reactor, the
medium composition should be the same as the feed
applied to the parent reactor. For comparison of the rate
to other conditions, it could be a consideration to use an
equal standard medium for each condition. However,
when the research question includes an investigation of
the effect of certain compounds on the SRB activity, the
compound of interest should be either added to or
removed from the media. If many tests of a similar nature
are to be executed, it is recommended to prepare a
sufficient volume of media in order to save on
preparation time. As described in Section 2.3.4.4,
suitable pH buffers can be added initially to the batch
incubation bottles to maintain the pH.
Depending on the objectives of the activity tests, the
COD composition and concentrations, as well as the
sulphate concentrations present in synthetic wastewater,
can vary since they are usually the subject of
investigation. Moreover, the concentrations may be
adjusted proportionally for the duration of the test and the
sludge concentration used. Usually, concentrations of up
to 50 and 100 mg RBCOD L-1 are used for tests with fullscale activated sludge samples and of up to 400 mg
RBCOD L-1 with lab-enriched cultures. Regardless of the
nature of the activity test, synthetic wastewater must
contain the required macro- and micro-elements (such as
potassium, magnesium, calcium, iron, zinc, cobalt,
among others) in sufficient amounts for SRB to avoid any
metabolic limitation that may jeopardize the outcomes of
the batch activity.
Postgate's medium is the most commonly used
medium for the growth of SRB and to carry out SRB
activity tests (Postgate, 1984). It consists of (in g L-1):
K2HPO4 (0.5), NH4Cl (1.0), Na2SO4 (1.2), FeSO4·7H2O
(0.5), COD (3.0), MgSO4·7H2O (2.0), yeast extract (1.0),
ascorbic acid (0.1), thioglycollic acid (0.1), and Na2SO4
(1.0). Sodium acetate (NaAcetate·3H2O) can be used as
the carbon source. However, several synthetic
wastewater recipes for performing SRB activity tests
have also been described in the literature depending on
the carbon source used (Villa-Gomez et al., 2011). van
den Brand et al. (2014a) recommended the following
media composition: NaAcetate·3H2O (486.2 mg L-1),
NaPropionate (147.1 mg L-1), K2HPO4 (37.8 mg L-1),
KH2PO4 (14.2 mg L-1), NH4Cl (382 mg L-1), MgCl2·6H2O
(93.4 mg L-1), CaCl2 (58.4 mg L-1), trace elements (1 mL
L-1), and an adequate sulphate source (for example:
MgSO4), resulting in a sulphate concentration of 500 mg
ACTIVATED SLUDGE ACTIVITY TESTS
L-1. For fermentative SRB activity tests, Villa-Gomez et
al. (2011) used the following media composition in batch
and continuous bioreactors: KH2PO4 (500 mg L-1),
NH4Cl (200 mg L-1), CaCl2·2H2O (2,500 mg L-1),
FeSO4·7H2O (50 mg L-1), MgSO4·7H2O (2,500 mg L-1)
and lactate as the electron donor. However, it is
noteworthy to remember that a combination of complex
carbon sources (lactate and VFA) and sulphate can be
used as long as the COD/SO42- ratio is < 0.67. However,
while carrying out these tests, blackening of the media
due to sulphate-reduction and sulphide production can
easily be noticed. Other examples of compounds that can
be used to introduce sulphate in synthetic wastewater are
seawater or sodium sulphate taking into account that the
COD/SO42- ratio is 0.67. Trace elements can be prepared
according to Lau et al. (2006). If a test requires a
particular COD/SO42-value in order to investigate the
effect of limited or excess sulphate levels, the media can
be simply adjusted by adding low amounts of sulphate or
more COD. When the media is to be prepared in advance,
follow the recommended separation of media containing
COD source and N source in order to avoid biomass
growth in the prepared media. The final colour of this
media is transparent. If required, the synthetic
wastewater can be concentrated, sterilized in an
autoclave (for 1 h at 110 °C) and used as a stock solution
if several tests will be performed in a defined period of
time. However, the solution must be discarded if any
precipitation or loss of transparency is observed.
For experiments performed with lab-enriched
cultures, it is recommended to perform the tests with the
same (synthetic) media that is used for the cultivation.
Alternatively and similar to the case when sludge from
full-scale plants is used, the effluent from the reactor can
be collected, filtered through rough pore size filters to
remove larger particles, and used to prepare the required
media with the desired carbon and sulphate
concentrations for the execution of the activity tests.
2.3.5 Analytical procedures
Analytical tests required for the SRB activity test can be
performed by following standardized and commonly
applied analytical protocols as described in Section
2.2.2.5 on EBPR.
However, for estimating the activity of SRB only for
a few selected parameters of interest, a modification of
standard techniques/protocols is required. Therefore, a
detailed description of the methods for COD and sulphide
determination is given below. The analysis for sulphate
63
is also included, as this parameter is of interest in SRB
activity tests. However, for each test, it is important to
realize that some compounds interfere with analytical
tests. For instance, COD determination is very sensitive
when chloride is present in the samples. The effect of
such compounds on the analytical tests should be verified
by carrying out the analysis without this compound.
Concerning sulphide concentration measurements,
this could be done by making a standard sulphide
concentration with a known concentration and a solution
with the suspected disturbing compound. Then two tests
can be performed in parallel. The first test should be
carried out using the standard solution and the second test
using the standard solution plus the suspected compound
at concentration levels that are usually expected in the
mixed liquor. If the sulphide concentration is the same in
both analytical tests, the effect of the potentially
disturbing compound is not significant in the tests at the
applied concentrations and therefore, the analytical test
should be considered as appropriate for analyses.
2.3.5.1 CODorganics and CODtotal
The method to measure COD has been extensively
described elsewhere (APHA et al., 2012), though some
important additions are required. It is usually
straightforward to determine the total COD of the sample
or only the soluble COD fraction. To measure the soluble
COD fraction, the sample should be filtered using a 0.45
μm filter. The sulphide also contributes to the COD
measurements, and often this contribution (interference)
is not desired. Therefore, two analyses should be
performed, named hereafter as CODorganics and CODtotal.
The CODtotal (also referred to as the total COD
concentration in a soluble state), in this particular case,
also includes sulphide, while in the CODorganics analysis,
interferences due to sulphide contribution can be
eliminated from the sample by applying a mathematical
correction and thus the COD concentration is only
associated with the content of organic matter present in
the sample. As recommended by Boyles (1997), prepare
a separate standard for sulphide and perform a COD test
on the standard, and apply a mathematical correction to
the result of CODtotal.
Another method to completely remove sulphide from
the sample is based on the precipitation based procedure
of Poinapen et al. (2009). A few drops of 10 M NaOH
solution are added to the sample, as an increase in pH will
induce precipitation. The H2S/HS- system has a pKs1 of
7.05 at 25 ˚C (Bjerrum et al., 1985) and at a pH of 10.0
most of the sulphide present in the sample is in the form
64
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
of HS- species with a very small proportion in the form of
S2- species (pKs2 = 12.92 at 25 ˚C). Neither HS- nor S2species can escape from the sample since both
compounds are very well dissociated. Subsequently, zinc
sulphate (ZnSO4) is added to precipitate with sulphide as
zinc sulphide (ZnS), which should be separated from the
sample by filtration. Finally, the measured
concentrations should be corrected for the dilution with
the ZnSO4 solution.
In the direct methylene-blue method, specific for
sulphide analyses, zinc acetate is used to ensure a higher
degree of sulphide fixation. Subsequently, the reagents
containing dimethyl-parafenyl-diamine and iron are used
to form methylene blue, resulting in a blue solution. The
intensity of the colour can be analysed photometrically to
quantify the sulphide concentration in the sample.
2.3.5.2 Sulphate
Gas-phase samples
Sulphate can be measured by various techniques, such as
gravimetric, ion chromatography, methylthymol blue
method and turbidimetric. These methods are described
elsewhere (APHA et al., 2012). The most commonly
used method is a Hach Lange kit, a turbidimetric-based
procedure or using a high performance liquid
chromatography (HPLC) equipped with an AS9-SC
Column and an ED 40 electrochemical detector (Dionex).
2.3.5.3 Sulphide
Depending on the pH, sulphide can be present in various
states. Therefore, it is important to perform a sulphide
test in both the liquid as well as the gas phase.
Liquid samples
When samples are taken for sulphide measurements, the
procedures applied are crucial as sulphide is easily lost
during sampling due to stripping or chemical reactions. It
is therefore important to solubilize the volatile fraction in
the liquid phase, for example using NaOH solution. For
the analysis of sulphide, two main procedures can be
used. The first procedure is referred to as the indirect
COD method, while the second procedure is the direct
methylene-blue method. Both these procedures will be
described in this section. 1 g of sulphide corresponds to
~2 g COD. The sulphide concentration could therefore be
expressed by its COD content.
This could be achieved by measuring the COD
concentration of the sample in terms of CODtotal and
CODorganics. The COD concentration should be analysed
using the standard method, according to the specific
treatment steps described previously (Section 2.3.5.1).
The sulphide concentration could then be calculated by
subtracting the CODorganics value from the CODtotal value,
and corrected for the COD/sulphide weight ratio. The
expression to calculate the sulphide content (SH2S in g L-1)
is:
SH2S = (CODtotal ‒ CODorganics ) / 2
Eq. 2.3.5
Where, 2 is the sulphide to COD ratio (w/w).
Gas-phase samples should be collected from the batch
incubation bottles using airtight glass syringes (for
example: Hamilton, USA). The most commonly
employed method to estimate gas-phase H2S
concentrations is using gas chromatography (GC). Li et
al. (2013) used a GC (Agilent 6890 N, USA) fitted with
a flame photometric detector (FPD) and DB-1701
capillary column (30 m × 0.32 mm × 0.25 μm, Hewlett
Packard, USA) to estimate H2S concentrations. The
temperature of the oven, injection, and detector were set
at 100 °C, 50 °C, and 200 °C, respectively, while nitrogen
was used as the carrier gas. H2S concentrations can also
be determined by titration using a standard potassium
iodide-iodate as the titrant and starch indicator (APHA et
al., 2012). Another quick and easy method to monitor the
H2S concentrations in wastewater treatment plants is the
use of gas detector tubes (Kitagawa, Japan). H2S
concentrations can also be measured using Jerome 631X Hydrogen Sulphide Analyser (Arizona Instruments,
USA). However, in cases where it is difficult to perform
the analysis immediately, Tedlar bags (Tedlar gas
sampling bags, Sigma-Aldrich) should be used for
collecting gas samples from continuously operated
reactors.
Temperature and pressure are two important factors
that need to be considered during the calibration of GC
for gas-phase H2S measurements. If S2- in liquid phase is
added to a closed system, the resulting gas-phase H2S
concentration (CH2S) in the system can be calculated as
follows:
CH2S = 22.4 + 106 ·
ρ · VL
T
760
+
+
MW · V 273
P
Eq. 2.3.6
Where, ρ is liquid density (g mL-1), VL is the volume
of liquid (mL), T is the temperature (K), P is the pressure
(torr), MW is the molecular weight (g mol-1) and V is the
volume of the closed system (L).
ACTIVATED SLUDGE ACTIVITY TESTS
2.3.6 SRB batch activity tests: preparation
This section describes not only the different steps but also
the apparatus characteristics and materials needed for the
execution of the SRB batch activity tests.
2.3.6.1 Apparatus
If the experiments are conducted using a reactor, the
description given in Section 2.2.3 on EBPR is applicable.
If the tests are conducted in serum bottles, the following
apparatus is required:
1. Serum vials, including a rubber stop and aluminium
crimps to secure an anaerobic environment.
2. A nitrogen gas supply.
3. A pH electrode.
4. A 2-way pH controller to add either HCl and/or
NaOH (alternatively, a one-way control for HCl
addition or manual pH control can be applied through
the manual addition of HCl and/or NaOH).
5. A thermometer (recommendable temperature
working range of 0 to 40 °C). Confirm that the
electrodes or meters (pH, thermometer) are calibrated
less than 24 h before execution of the batch activity
tests in accordance with guidelines and
recommendations from suppliers.
6. A room or incubator to control the temperature at the
desired temperature.
7. A shaker in which the rpm can be set up to 300 rpm.
8. A pipette and tips to measure the exact volume of the
samples taken.
9. Syringes and needles to take the sample from the
serum vials by making a vacuum.
10. Filters with a pore size of 0.45 μm for sample filtering
and for preservation purpose.
11. Additionally, all the materials required for analysis.
These are elaborated in Section 2.3.5.
2.3.6.2 Materials
1. Two graded cylinders of 1 or 2 L volume (depending
upon the sludge volume used) to hold the sample and
wash it, if required.
2. At least 2 plastic syringes (preferably of 20 mL or at
least of 10 mL volume) for the collection and
determination of soluble compounds (after filtration).
3. At least 3 plastic syringes (preferably of 20 mL
volume) for the collection of solids, particulate or
intracellular compounds (without filtration).
4. At least 3 glass syringes (preferably of 5 mL volume)
for the collection of gas-phase samples.
65
5. 0.45 μm pore size filters. Preferably not of cellulose
acetate because they may release some traces of
cellulose or acetate into the collected water samples.
Consider having at least twice as many filters as the
number of samples that need to be filtered for the
determination of soluble compounds.
6. Transparent plastic cups with a volume of 10 or 20
mL to collect the samples for the determination of
soluble compounds (e.g. soluble COD, acetate,
propionate, sulphate, and sulphide).
7. Transparent plastic cups with a volume of 10 or 20
mL to collect the samples for the determination of
MLSS and MLVSS. Consider the collection of these
samples in triplicates due to the variability of the
analytical technique.
8. A plastic box or dry ice box filled with ice with the
required volume to temporarily store (for up to 1-2 h
after the conclusion of the batch activity test) the
plastic cups and plastic tubes for centrifugation after
the collection of the samples.
9. Plastic gloves and safety glasses.
10. Pasteur or plastic pipettes for HCl and/or NaOH
addition (when the pH is controlled manually).
11. Metallic lab clips or clamps to close the tubing used
as a sampling port when samples are not collected
from the reactor/fermenter.
2.3.6.3 Media
• Real wastewater
See Section 2.2.3.3 on EBPR.
• Synthetic media or substrate
If tests require synthetic wastewater, depending on
the type of the tests, the synthetic influent media
could contain a mixture of carbon and sulphate
sources plus relevant (macro- and micro-) nutrients.
Generally, they can be mixed all together in the same
media or split in two solutions (i) COD source, and
(ii) SO42- and N source plus the nutrient solution).
a. COD source solution: It must be composed of
RBCOD, preferably of VFA such as acetate or
propionate, depending on the nature or goal of the
test and the corresponding test objective (research
questions). For anaerobic batch activity tests, the
concentration needs to be adjusted to ensure that
the COD is consumed within the duration of the
test. For batch activity tests performed with
activated sludge from a full-scale plant, usually
COD concentrations not greater than 100 mg L-1
are recommended. For lab-scale mixed liquor
sludge samples, the COD concentrations can be
66
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
as high as the influent COD concentration of the
lab-scale system (and even sometimes 2 to 3
times higher). An example of a COD solution is
NaAcetate·3H2O (486.2 mg L-1) or Na-Propionate
(147.1 mg L-1).
b. Sulphate source solution: Sulphate can be added
as sodium sulphate. However, in some cases and
especially in the case of saline wastewater
treatment, sulphate is accompanied by salinity.
To mimic those conditions, it is also possible to
add sea salt (containing sulphate). The sulphate
concentrations can be adjusted, as desired,
depending on the purpose of the experiment. The
nutrient solution should contain all the essential
macro- (ammonium, magnesium, calcium,
potassium) and micro-nutrients (iron, boron,
copper, manganese, molybdate, zinc, iodine,
cobalt) to ensure that SRB metabolism is not
limited. This can lead to the wrong interpretation
of results and in extreme cases, a near complete
failure of the test. Thus, one must make sure that
all the required compounds are added to the
solution and in the correct amounts. The media
composition is recommended in Section 2.3.4.8.
c. Washing media: If the sludge sample must be
washed to remove the presence of an undesirable
compound (which may be even inhibitory or
toxic), prepare a fresh nutrient solution to wash
the sludge containing the medium composition
described above, without VFA. The washing
process can be repeated two or three times
following a similar procedure likewise in Section
2.2.3.5. Thereafter, the necessary preparation
steps of the batch activity tests can be performed.
d. Preparing acid and base solutions: see Section
2.2.3.3 on EBPR.
2.3.6.4 Material preparation
•
•
•
•
If analyses are not outsourced to specialized
analytical labs, the required stock and working solutions
to carry out the determination of the analytical
parameters of interest must be also prepared in
accordance with standard methods (APHA et al., 2012)
and the corresponding protocols.
For information on the number of samples, see
Section 2.2.3.4 on EBPR batch activity tests.
Frequency of sample collection:
a. If the maximum specific kinetic rates must be
determined (e.g. maximum specific acetate or
propionate uptake rate or maximum sulphatereduction rate), then increase the frequency of
sampling during the initial few hours or days of
incubation. In general, samples should be
collected once every 5 min during the first 30-40
min of duration of the batch activity test, similar
to the method described for EBPR tests.
b. To ascertain the stoichiometric ratios and not the
kinetic ones (e.g. anaerobic SO42- reduction/HAc
uptake ratio), then the samples can be collected
only during the beginning and at the end of each
phase to determine the conversions of interest
within the selected period (phase).
To increase the reliability of data collected and to
determine the initial conditions of the sludge, it is
strongly recommended to collect a series of samples
before any media is added. Carefully define the
maximum and minimum working volumes of the
reactor, as described in Section 2.2.2 on EBPR.
Preparation of sample cups can be performed as
described in Section 2.2.3.2 on EBPR. A simple
working plan created in a spreadsheet (Section 2.3.9)
can be rather useful to perform and keep track of the
sample collection frequencies. Furthermore, this
spreadsheet can be used to maintain a database of the
different batch tests carried out. Organize all the
required material within a relatively close radius of
action, around the batch setup, so that delay in
handling and preparing the samples can be avoided.
Calibrate all the meters (pH and thermometer) less
than 24 h prior to the execution of the tests and store
them in proper solutions until the execution of the
tests, following the particular recommendations of
the corresponding manufacturer or supplier and
confirm that their readings are reliable before the start
of the tests. Samples must be properly stored and
preserved until they are analysed (Table 2.3.6).
Table 2.3.6 Recommended sample storage and preservation procedure for the determination of SRB activity tests.
Parameter Material of
Method of preservation
sample container
Sulphate Plastic or glass Filter immediately after sample collection thorough 0.45
μm pore-sized filter and cool to 1-5 °C, or freeze to -20 °C.
Sulphide Plastic or glass Immediately solubilize in a drop of 1M NaOH.
Maximum recommended time between sampling, preservation
procedure and analyses
24 h for samples stored at 1-5 °C; up to 1 month for frozen samples.
Recommended to perform these measurements as quickly as possible.
ACTIVATED SLUDGE ACTIVITY TESTS
2.3.6.5 Mixed liquor preparation
These procedures consider that batch activity tests can be
performed as soon as possible after collection of samples
from full- or lab-scale systems or, in the worst case
scenario, within 24 h after collection. Performing the
batch tests 24 h after the collection of mixed liquor sludge
samples is not recommended because the SRB culture
can undergo potential biochemical changes during
handling (unless the exposure time after collection is of
particular interest for performing these tests). Bearing in
mind the previous comments, the following three
protocols are recommended to prepare the mixed liquor
samples for conducting batch activity tests:
1. If batch activity tests can be performed in less than 1
h after collection of the sludge sample and if the
sludge sample does not need to be washed:
a. Adjust the temperature of the batch reactor where
the tests will take place to the target temperature
of the study.
b. Collect the sample at a wastewater treatment
plant (WWTP) from one of the following:
- From the raw influent.
- At the end of the aerobic tank or stage.
- At the outlet of the primary or secondary
sludge thickeners or anaerobic digesters.
- Take sludge from a laboratory reactor.
c. Transfer the sludge sample to the reactor or serum
bottles where the batch activity tests will take
place.
d. Start a gentle mixing (50-100 rpm in the case of
the reactor) and follow up the temperature of the
sludge sample by placing a thermometer inside
the reactor or serum bottle (if the setup does not
have a built-in thermometer).
e. Maintain mixing conditions until the sludge has
reached the target temperature of the study.
f. Maintain the anaerobic conditions by using an
airtight reactor and sparge N2 gas to avoid/reduce
oxygen intrusion.
g. If nitrate is detected in the aerobic samples
collected to perform these tests then follow
instructions presented in Section 2.2.2.3 on
EBPR. Alternatively, for nitrate concentrations
greater than 10 mg NO3--N L-1, 9 mg COD could
be added per every mg of nitrate detected (9 mg
COD mg NO3--N-1). As soon as nitrate is no
longer observed, the sludge can be immediately
used to conduct the anaerobic batch activity tests.
Nitrate or nitrite detection strips (Sigma-Aldrich)
can be used for a quick estimation of the presence
of these compounds.
67
2. If batch activity tests can be performed in less than 1
h after collection of the sludge sample but the mixed
liquor sludge sample needs to be washed, see
washing procedures described in Section 2.2.3.5.
a. Transfer the 'washed' sludge sample to the reactor
or fermenter where the batch activity tests will be
performed.
b. Start a gentle mixing (50-100 rpm in the case of
the reactor) and follow up the temperature of the
sludge sample by placing a thermometer inside
the reactor (if the set-up does not have a built-in
thermometer).
c. Maintain mixing conditions until the sludge has
been exposed to the target temperature of study
for at least 30 min.
3. In some cases, due to location and distance issues, the
tests cannot be performed in less than 1 or 2 h after
collection (but within 24 h):
a. Adjust the temperature of the batch reactor to the
target temperature of the study.
b. Keep the sludge sample cold until the
commencement of the test (e.g. by placing the
bucket or jerry can in a refrigerator at 4 °C).
c. Prior to the test, take out the sludge sample from
the refrigerator, cool box or cold room.
d. Mix the contents gently in order to obtain a
homogenous and representative sample with a
similar MLSS concentration as in the original labor scale system from where it was originally
collected.
e. If the sample needs to be washed, wash the sludge
in a mineral solution (see Section 2.3.6.3 on
working solutions).
f. Transfer the washed sludge to the reactor or
fermenter where the batch activity test will take
place.
g. Start to flush N2 gas in the sludge sample while
mixing gently to remove any dissolved oxygen.
h. Follow the temperature of the sludge sample by
placing a thermometer inside the reactor (if the
setup does not have a built-in thermometer).
i. Keep flushing N2 gas and mix the contents for at
least 1 h (maximum 2 h), but ensure that the
sludge is exposed to the target temperature of the
study for at least 30 min.
j. After the procedure of adjusting the temperature,
if nitrate is detected (see Section 2.3.6.3), add a
solution containing readily biodegradable COD at
a 9 mg COD mg NO3—N-1 ratio. As soon as nitrate
is no longer observed, the sludge can be
immediately used to start and conduct the
anaerobic batch activity tests.
68
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.3.6.6 Sample collection and treatment
For the purpose of this activity test, three types of
samples are required: (i) a filtered sample, (ii) a filtered
sample with NaOH and (iii) a biomass sample. Table
2.3.7 provides a list of analyses corresponding to
different sample collection and treatment steps.
Table 2.3.7 List of analyses performed on each sample.
Sample
Filtered sample
Filtered sample with NaOH
Biomass sample
Parameter
CODtotal; CODorganics; Sulphate
Sulphide
MLVSS; MLSS
• Filtered sample
The filtered sample is used for the analyses of CODtotal,
CODorganics and sulphate. A well-mixed sample is taken
from the serum vial with a needle and a syringe. The
needle is injected into the rubber stopper. Then the serum
bottle is turned upside down, so that the needle is
immersed in the liquid phase. The syringe is pulled in
order to create a vacuum and collect the sample.
Subsequently, the sample is directly filtered using a 0.45
μm filter, to avoid further conversions. Thereafter, the
CODtotal, CODorganics and sulphate concentration are
measured according the protocol described in Section
2.3.5.
• Filtered sample with NaOH
The filtered sample in which a few drops of NaOH are
added is used for the analyses of the sulphide
concentration. The sample should be transferred into a
tube containing three drops of 1 M NaOH directly after
sampling. This is done by using a needle and a syringe.
This needle is pierced into the rubber stopper (septa).
Then the serum bottle is turned upside down, such that
the needle is immersed in the liquid phase. The syringe is
pulled back in order to create a vacuum and collect the
sample. The sooner the sample is mixed with NaOH, the
less sulphide is lost during sampling. Thereafter, the
sample is filtered using a 0.45 μm pore-size filter and
analysed as described in Section 2.3.5. This analysis
should be performed directly after collecting the sample,
to avoid sulphide losses.
immediately after collecting the sample from the glass
bottles in order to avoid losses.
• Biomass sample
In order to estimate MLVSS and MLSS concentrations,
a well-mixed sample from the reactor should be
collected. The easiest method is to use a syringe and
needle arrangement. Ensure that you know the exact
volume taken from the reactor. The MLVSS and MLSS
should be analysed in triplicate as described elsewhere
(APHA et al., 2012). The frequency of analysing COD,
sulphate and sulphide depends on the biomass activity.
After data processing, a linear line (as presented in Figure
2.3.8) is required for proper analyses. The more data
points present on this linear line, the better, but a balance
between investments (in money and time) suggests that
four points on the linear line are already sufficient. The
safest option is to take several samples during the first
hour, once every 10 or 15 min. The best way to deal with
this is to perform extensive sampling in the first tests (as
a trial), and based on these results, a design can be made
to ascertain how much data points are actually required
to perform subsequent tests.
2.3.7 Batch activity tests: execution
This section describes how a batch activity test of SRB
should be executed. The description consists of a material
and chemical list, protocol to prepare media, sampling
scheme (time frame), sample collection, and analysis.
This section concludes with a step-by-step approach to
performing sulphate-reducing bacteria activity tests. The
material and chemical list specific for the analytical tests,
as well as how the tests should be conducted, is described
in Section 2.3.6.
The general advice as provided in Section 2.2.4 on
EBPR should be following accordingly for the SRB
activity tests. Table 2.3.8 gives examples of typical tests
performed to ascertain SRB activities.
Table 2.3.8 Typical tests for SRB activity analyses.
Test code
SRB.ANA.1
• Gas sample
Gas-phase samples should be collected from the batch
incubation bottles using air-tight glass syringes. A
calibration plot should be prepared in the GC using
standard H2S gas cylinders (example: 25 to 1,000 ppm).
Gas-phase analysis should also be performed
SRB.ANA.2
Redox
Short description and purpose
Anaerobic Performed with real wastewater, for instance,
the raw influent, to determine the exact rate of
SRB under that particular condition.
Anaerobic Conducted using a defined standard media (see
Section 2.3.4.8), to compare the sulphatereduction rate from SRB under different
conditions.
ACTIVATED SLUDGE ACTIVITY TESTS
Test SRB.ANA.1 Anaerobic SRB activity test
The first step of conducting a SRB activity test is to
ensure the correct composition of medium and following
a correctly designed experimental schedule (Section
2.3.8.2). Both steps heavily rely on the goal and scope of
one's research, but also on the origin of the SRB sludge
(enriched culture, mixed culture or culture collected from
a full-scale or lab-scale system). The required volume of
the sludge sample depends on the objective of the study
and the research questions. It should be ensured that the
biomass and liquid/gas phase concentrations are
comparable as much as possible to the original situation,
unless required otherwise. After the experimental design,
the steps shown below should be followed. The same
approach is also valid for SRB activity tests in reactors.
1. Collect the items on the material and chemical lists.
2. Prepare the required stock solutions for conducting
analytical tests.
3. Prepare the media stock solutions.
4. Fill the serum vials with media and correct the pH, if
necessary, with HCl and/or NaOH.
5. Transfer the SRB sludge into the serum bottles
(preserving the anaerobic conditions as much as
possible, for example, by avoiding turbulence).
6. Close the serum bottles with a rubber stopper and
aluminium crimp.
7. Flush both the headspace and the liquid phase in the
serum bottles with N2 gas. This is done by injecting a
needle with the gas supply and placing another needle
slightly above the liquid phase from which the
overpressure can escape. This latter needle should not
touch the liquid phase; otherwise, the sludge will start
to escape. For 80-110 mL bottles, 1 min flushing is
enough to obtain the required anaerobic conditions,
but this could be checked by the addition of resazurin
(dying colour). The media becomes colourless after
becoming anaerobic. Remove both needles from the
septa at exactly the same time. The bottles should
now be anaerobic.
8. Now take a sample at time t = 0 and handle this and
other samples immediately. The sample is collected
using a syringe. Then turn the bottle such that the
needle touches the liquid (avoid taking out biomass).
Pull the syringe back to collect the sample.
9. Collect the sulphide sample in 1 drop of 1 M NaOH,
and the rest can be used for COD and sulphate
analyses. Sulphide samples should be solubilized in
NaOH and immediately analysed.
10. Follow the time scheme of the sampling.
69
11. After finishing the time scheme, open the serum
bottles and collect the biomass sample and perform
MLVSS and MLSS analyses.
12. If the experiments cannot afford SRB sludge losses
and it is desired to put the sludge back into the parent
reactor, wash it first three times with the media used
to feed the parent reactor. If the media composition is
not known, use demineralised water for washing.
13. Then the data should be analysed using the procedure
described in Section 2.3.8.
14. Note: if the reactor is used for batch activity tests, a
similar procedure as described for EBPR anaerobic
tests should be followed (see Section 2.2.4).
Test SRB.ANA.2 Anaerobic SRB activity test
A similar approach should be followed for SRB.ANA.2
using synthetic media (Section 2.3.6.3) to estimate the
SRB activity.
2.3.8 Data analysis
2.3.8.1 Mass balances and calculations
The experimental data can be used to calculate the
sulphate-reducing activity of SRB (Section 2.3.8.2).
Sulphide formation is a result of SRB activity and the
variability of rate coefficients depends on the carbon
source, initial sulphate concentrations and the presence
of specific SRB genera. The additional COD, sulphate
and sulphide concentration measurements taken at time t
= 0 and at the end of the measurement (effluent) should
be used to check the mass balances. Rate determination
can only be reliable with the correct mass balances in
place. The mass balance equations for COD and sulphur
compounds can be represented as follows:
COD balance
CODorganics,in + SH2S,in = CODorganic,out + SH2S,out Eq. 2.3.6
Sulphur balance
SSO4,in + SH2S,in = SSO4,out + SH2S,out
Eq. 2.3.7
Also it is possible to calculate an electron balance.
To calculate the biological sulphate-reduction rate
(only when the mass balances are correct), the following
calculations should be executed:
a. Use a standard calibration curve to convert the
measured optical density values (OD675) from the
spectrophotometer to concentration units (mg L-1).
70
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
b. Convert the sulphide concentration from mg L-1 into
mM (if desired). The molar mass of completely
dissolved sulphide (S2-) is 34 g mol-1.
c. Take the average value of the triplicate.
d. To obtain a profile indicating the amount of sulphide
produced, correct for the sulphide concentration
present at time t = 0 min (by subtracting this value
from all the values).
e. Determine the slope from the first samples, in which
a linear line is expected.
f. Divide the slope value by the amount of biomass
present (g VSS), to determine the rate in mmol SO42g VSS-1 h-1). The formation of 1 mol sulphide is
related to the removal of 1 mol sulphate. Also a
correction from molarity to mol is required.
wastewater treatment, and sometimes with low sulphate
concentration in the influent, resulting in rates that were
significantly lower than the ones observed by van den
Brand (2014a; 2014b; 2015).
Table 2.3.9 Sulphate-reduction rates (qSRB, SO4,An) reported in the literature
from sludge taken from reactors after long-term operation.
RBC
Full
qSRB,SO4,An
(mg SO42- g VSS-1 h-1)
1-11
CAS
Full
0-3.1
Lens et al. (1995)
CAS
Laboratory
1.9
Lens et al. (1995)
ANS
Full
6-7.3
Lens et al. (1995)
ANS
Laboratory
11.6
Lens et al. (1995)
2.3.8.2 Data discussion and interpretation
ANS
Laboratory
100-220
The results of COD, sulphate and sulphide analyses of
the samples taken during the start and the end of the test
can be used to check the mass balances. As sulphide is
easily lost during sampling, a mass balance of 100 % is
not easy to obtain. Therefore, a mass balance fit of 95 %
is considered to be satisfactory in order to continue data
analyses. If only sulphide formation in a batch test is
analysed to measure the biological sulphate-reduction
rate by SRB, it is important to perform the test at least in
triplicate. It is also very important to check the mass
balances, which is possible by executing COD and
sulphate analyses at time t = 0 and at the end of the test.
The methylene-blue method is a very sensitive and
reliable technique. Therefore, when the mass balance fit
is around 95 %, the sulphide accumulation profile can be
the basis for the calculation of the reduction rate. It is also
possible from this type of analyses to determine the
characteristic conversion parameters, such as the
maximal conversion rate. To measure the maximal
conversion rate, it is crucial that all compounds are
present in excess. Then, the maximal conversion rate can
be determined as described in Sections 2.3.9 and 2.2.5.2.
The Lineweaver-Burkplot method can be used to
determine the characteristic parameters. This method is
described in detail by Nelson and Cox (2005).
RBC: Rotating Biological Contactor
CAS: Conventional Activated Sludge
ANS: Anaerobic Sludge
References
Lens et al. (1995)
van den Brand et al.
(2014a; 2014b; 2015)
2.3.9 Example
Table 2.3.10 shows an example of worksheet that can be
used for different SRB activity tests. The example
concerns the tests solely based on the production of
sulphide and COD, sulphate and VSS measurements
taken during the beginning and end of the experiments.
Figure 2.3.8 is an example of a test performed with
synthetic wastewater that also includes the analysis of
VFA (acetate and propionate).
200
Sulphate, Sulphide (mg L-1), CODorganics (mg L-1)
There are only limited results available for an SRB
activity test, specifically developed for domestic
wastewater treatment. Some available literature results
are presented in Table 2.3.9 and evidently a wide range
of sulphate-reduction rates have been reported depending
on the operating conditions that facilitated SRB activity.
Although the SRB are strictly anaerobic microbes, some
studies (e.g. Lens et al., 1995) are based on aerobic
System Scale
180
160
140
120
100
80
60
40
20
0
0
60
120
180
240
300
360
Time (h)
Figure 2.3.8 Example of a graph depicting sulphide (●), sulphate (●)
and CODorganics (●) concentrations during the SRB.ANA.2 activity test from
which the sulphate-reduction rate can be calculated.
ACTIVATED SLUDGE ACTIVITY TESTS
71
For this test, it is suggested to perform several (at
least 10) COD and sulphate measurements, and VSS/TSS
measurements from the sludge at the end of the test, as
described in Section 2.3.7.
SRB activity tests could also be used to investigate
the effect of compounds on the sulphate-reduction rate.
Figure 2.3.9 presents an example of a study comparing
different nitrogen compounds (ammonium, nitrate and
nitrite) at different N concentrations as performed by van
den Brand et al. (2015). By calculating the sulphatereduction rate under specific conditions as performed in
Figure 2.3.9, and by presenting all the SRB rate results in
one figure, the effect of nitrogen compounds on the
sulphate-reduction rate can be analysed. In this case error
bars are included to show the accuracy of the
measurements. The sulphate-reduction rate decreased in
time with an increased nitrogen concentration (Figure
2.3.9).
Using the protocol described in Section 2.3.5, a graph
is obtained, from which the slope can be determined. A
typical profile obtained from an SRB activity test is
presented in Figure 2.3.8. The slope calculated in this
example is 225 mg SO42- h-1, which corresponds to 2.34
mmol SO42- h-1. The amount of biomass in the reaction
vessel was 4.5 g MLVSS. The working volume of the
batch reactor was 2.5 L. The actual biological sulphatereduction rate was subsequently calculated as follows:
qSRB,SO4,An =
2.34
= 0.53 mol SO4 g VSS-1 h-1
4.5
Eq.2.3.8
Table 2.3.10 A typical example of a worksheet for an SRB activity test (type SRB.ANA.2). Note that when more information regarding the kinetics is desired,
sulphate and CODorganics analysis should be performed during all the selected times. However, instead of CODorganics, the actual acetate and propionate
concentrations should be measured.
Anaerobic sulphate reduction batch tests
Date:
Description:
Test No.:
Duration
Code: SRB.ANA.2
Wednesday, 09.10.2015 10:00 h
Test at 10 °C, pH 7, synthetic substrate and enriched SRB culture
1 of 3
6.0 h (360 min)
-1
Synthetic: acetate and propionate (total 400 mg L )
Serum vial using needle
SRB.ANA.2(1.1-1.11)
80 mL
(20 mL for MLVSS, 6 mL for other samples)
2.5 L source sludge, 80 mL serum vial
Substrate:
Sampling point:
Samples No.:
Total sample volume:
Reactor volume:
Sampling schedule
Time (min)
Time (h)
Sample No.
Parameter
-1
COD (mg L )
0
0.00
2
1
136
95
59
39
25
9
8
8
8
8
18
40
58
74
83
81
83
83
84
85
195
176
153
137
124
107
102
96
96
396
-1
-1
Sulphate (mg L )
30
0.50
4
4. Take sample for COD (acetate and propionate: CODorganics ) in the feed.
5. Transfer 100 mL of sludge to serum vial.
6. Close serum vial with a rubber stop and aluminium cap.
7. Flush serum vial headspace with N2 gas.
8. Place vial on mixing plate and take sample at t = 0.
9. Take other samples following the sampling scheme below.
10. After last sample is taken, measure biomass concentration of the remaining vial content.
11. Verify that all systems are swtiched off.
-20
-0.33
1
Sulphide (mg L )
15
0.25
3
Experimental procedure in short:
1. Confirm availability of sampling material and required equipment.
2. Confirm calibration and functionality of systems, meters and sensors.
3. Fill serum vial with media and correct for pH.
60
90
1.00
1.50
5
6
ANAEROBIC PHASE
120
2.00
7
180
3.00
8
240
4.00
9
-1
MLSS and MLVSS (mg L )
1
MLSS and MLVSS measurements
Sampling point
Cup No.
End anaerobic phase
1
2
3
-1
MLSS (mg L )
-1
MLVSS (mg L )
Ratio
-1
Ash (mg L )
360
6.00
11
91
See table
Average value of the COD concentration present in the synthetic substrate prior the start of the test
Biomass composition
Sampling point
300
5.00
10
W1
0.08835
0.08835
0.08834
End anaerobic phase
W2
0.10741
0.10759
0.10683
W3
0.08849
0.09018
0.08940
W2-W1
0.01906
0.01924
0.01849
W2-W3
0.01892
0.01742
0.01742
Average
MLSS
1,906
1,924
1,849
1,893
Note:
-1
1893
Acetate (CH2 O)
30.03 mg C-mmol
1792
Propionate (CH2 O0.67 )
24.69 mg C-mmol
-1
0.95
Sulphide (H2 S)
34.08 mg S-mmol
-1
101
Sulphate (SO4 )
96.06 mg S-mmol
-1
2-
MLVSS
1,892
1,742
1,742
1,792
Ratio
0.99
0.91
0.94
0.95
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
qSRB,SO4,An (mmol SO 2- g VSS-1 h-1)
4
72
1.8
2.3.10 Practical recommendations
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0
200
400
600
800
1,000
Concentration N (mg L-1)
Figure 2.3.9 Example of a graph that shows the biological sulphatereduction rate (qSRB,SO4,An) in the presence of ammonium (●), nitrate (●)
or nitrite (●) as N source in the feed, respectively.
It is also possible to study the relation between the
sulphide concentration and the actual sulphate-reduction
rate. Therefore, it is important to determine the moving
average along the profile, in order to know the rate for
the actual average sulphide concentration present in the
sample. The moving average is a result of the average of
three rates. These rates correspond to the sulphide
concentration at which the moving average is calculated,
and one rate at higher and lower concentrations. Figure
2.3.10 is an example of a typical relation between the
sulphate-reduction rate and the actual sulphide
concentration, based on the moving average method.
This type of graph can also be prepared for the actual
sulphate or COD concentration.
Volumetric sulphate-reduction rates and SRB activities
can be estimated by performing experiments under
controlled laboratory conditions. Figure 2.3.11 is an
example of a suspended growth bioreactor, fitted with
adequate monitoring devices for pH and temperature
control. Anew, such reactor configurations will facilitate
the ease of collecting both liquid and gas phase samples.
If an aerobic test is conducted to ascertain the influence
of oxygen on SRB activity, then the same tests should be
carried out, except for headspace flushing with N2-gas. If
a toxicity test is desired, then the tests using synthetic
wastewater should be performed and the toxic parameter
added to the experimental design. The test should be
repeated by using the same sludge that was exposed to
that toxic compound as the inoculum to investigate the
ability of the sludge to recover from toxic stress. From a
microbial perspective, if one wishes to determine the
time-dependent development of the SRB community
structure within the sludge, molecular biology tools like
fluorescence in situ hybridization (FISH) and denaturing
gradient gel electrophoresis (DGGE) of PCR-amplified
16S ribosomal DNA (rDNA) can be used (see Chapter
8).
qSRB,SO4,An (mmol SO42- g VSS-1 h-1)
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0
25
50
75
100
125
150
175
200
Sulphide concentration (mg L-1)
Figure 2.3.10 The relation between the sulphate-reduction rate
(qSRB,SO4,An) and the actual sulphide concentration, calculated by moving
average method.
Figure 2.3.11 Collection of a liquid sample to determine the activity of
sulphate-reducing bacteria (SRB). Note the characteristic black colour of
the enriched SRB biomass in the reactor (photo: KWR Watercycle Research
Institute, 2014).
ACTIVATED SLUDGE ACTIVITY TESTS
2.4 BIOLOGICAL NITROGEN REMOVAL
2.4.1 Process description
The need to remove nitrogen from wastewaters arises
from its potential toxic effect on aquatic life in receiving
water bodies as free ammonia (NH3), its effect on the
nitrogenous oxygen demand and on its role as a nutrient
in enhancing eutrophication especially in marine
environments (Metcalf and Eddy, 2003). Nitrogen is
mainly present in wastewaters in its reduced form as
ammonium (NH4) and can be removed by different
physicochemical and biological processes. The selection
of the best alternative is commonly based on cost
effectiveness. In general, physicochemical methods such
as ammonium air-stripping, breakpoint chlorination and
selective ion exchange are characterized by higher
operational costs that are considered economically
feasible only when the ammonium concentrations are
higher than 5 g N L-1 (Mulder, 2003). In wastewater
streams containing less than 5 g N L-1, as in several
industrial effluents and most municipal wastewaters
(where typical nitrogen concentrations are usually lower
than 100 mg N L-1), biological nitrogen removal
processes are usually preferred due to lower operational
costs. Different biological processes, and combinations
thereof, can be applied involving the relative metabolic
pathways. Among the resulting alternatives, the
wastewater characteristics in terms of the influent
COD/N ratio will provide guidance in determining the
most suitable biological processes for nitrogen removal.
Arguably, three influent COD/N ratio ranges can be
distinguished:
(i) For high influent COD/N ratios (> 20 g COD g N-1),
the nitrogen requirements or nitrogen assimilation
of heterotrophic bacteria (ordinary heterotrophic
organisms: XOHO) for biomass synthesis during
COD (organic matter) removal is usually sufficient
to achieve the required nitrogen concentration in
the effluent.
(ii) For influent COD/N ratios comprised between 5
and 20 g COD g N-1, the combination of nitrogen
assimilation for microbial growth and the
application of conventional nitrification and
heterotrophic denitrification processes can be
applied.
(iii) For COD/N ratios lower than 5 g COD g N-1,
conventional nitrification and heterotrophic
denitrification processes can hardly reach
satisfactory nitrogen removal levels. In particular,
the heterotrophic denitrification process will be
73
limited by the lack of organic matter and an
additional carbon source needs to be externally
dosed. Thus, non-conventional nitrogen removal
processes performing via the so-called ‘nitriteroute’ are more suitable for nitrogen removal
because of their lower (or even absent) COD
requirements. For this reason, besides other
technical and economic considerations, processes
such as partial nitritation-denitritation (also known
as nitrite shunt) or partial nitrification-anammox
(PNA) are nowadays state-of-the-art biological
nitrogen removal processes for wastewaters with
low influent COD/N ratio. Due to the relatively
lower growth rate of anaerobic ammonium
oxidizing bacteria, the anammox (anaerobic
ammonium oxidation) process is currently mainly
applied for the treatment of warm streams (> 25 °C)
such as the supernatant from the anaerobic sludge
digestion systems in municipal and industrial
WWTPs.
Overall, depending on the wastewater characteristics
and local conditions, the removal of nitrogen from
wastewater can be performed by several technologies and
combinations thereof. In spite of the different
technologies and N-removal processes, in practically all
of them ammonium is first (fully or partially) oxidized to
nitrite (nitritation) or to nitrate (nitrification) and then the
oxidized form of nitrogen is reduced to dinitrogen gas
which is released into the atmosphere via either
denitrification (from nitrate to N2), denitritation (from
nitrite to N2) or anammox (from nitrite and ammonium to
N2) processes.
The nitrogen removal efficiency of the biological
processes depends on an adequate balance between the
activities of the different microbial groups. In this regard,
the execution of batch activity tests to assess and
determine the stoichiometry and kinetic rates of these
biological conversion processes represents a useful tool
for monitoring and controlling the nitrogen removal
processes. Before describing in detail the batch test
methodologies and procedures, the biological processes
of nitrification, denitrification and anaerobic ammonium
oxidation are briefly presented.
For a deeper understanding of different process
configurations, operational conditions and factors
affecting each process as well as the metabolism involved
in the biological nitrogen removal cycle, the reader is
referred to standard textbooks (e.g. Henze et al., 2008;
Grady et al., 2011).
74
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
metabolic (catabolism plus anabolism) macro-chemical
reaction equation becomes:
2.4.1.1 Nitrification
Nitrification is the name given to the production of nitrate
by Schlœsing and Müntz (1877) who first recognized the
biological nature of the process. One decade later,
Winogradsky (1890) isolated for the first time an
ammonium-oxidizing bacterium showing that specific
groups of bacteria were responsible for nitrification. The
immense work of bacterial cultivation performed by
Winogradsky (1892) led to the isolation of Nitrosomonas
europea and Nitrobacter, showing that the production of
nitrate from the oxidation of ammonium was actually
divided into two distinct microbiological processes
performed by two phylogenetically independent groups
of chemolithoautotrophic aerobic bacteria: ammoniumoxidizing organisms (XAOO) producing nitrite and nitriteoxidizing organisms (XNOO) producing nitrate.
Ammonium oxidation is mostly performed by
chemolithoautotrophic ammonium-oxidizing bacteria
that use ammonium as their energy and nitrogen source
and inorganic carbon as the carbon source. Ammonium
oxidation has been reported to be performed also by
certain Archea (XAOA, Könneke et al., 2005) and by
heterotrophic bacteria (van Niel et al., 1993), showing
that the diversity of ammonium-oxidizing organisms is
larger than previously assumed. However, due to the
dominance of the chemolithoautotrophic pathway
performed by XAOO in wastewater treatment systems, this
is the ammonium oxidation process considered in this
chapter.
At the microbial process level, the oxidation of
ammonium to nitrite proceeds through the formation of
hydroxylamine (NH2OH) as the intermediate via the
enzyme ammonia monooxygenase (AMO):
NH+4 + O2 + H+ + 2e- → NH2 OH + H2 O
Eq. 2.4.1
Hydroxylamine is then further oxidized to nitrite by
the enzyme hydroxylamine oxidoreductase (HAO):
NH2 OH + H2 O → NO-2 + 5H+ + 4e-
Eq. 2.4.2
The combination of the two redox processes is known
as nitritation:
NH+4 + 1.5O2 → NO-2 + H2 O + 2H+
Eq. 2.4.3
This equation represents the catabolic macrochemical reaction equation of the nitritation process.
When the anabolism is also considered, then the
NH+4 + 1.383O2 + 0.09HCO-3 → 0.982NO-2 +
1.036H2 O + 0.018C5 H7 O2 N + 1.892H+
Eq. 2.4.4
As presented in the equation above, about 2 moles of
protons are produced per mole of ammonium oxidized.
The carbonate system is usually the pH buffer available
in the wastewater that neutralizes the production of
protons through CO2 stripping. When the carbonate
buffer, usually measured in terms of alkalinity as calcium
carbonate equivalents (CaCO3, meq L-1), is not available
or is insufficient in the wastewater (e.g. in the case of
municipal wastewater, alkalinity lower than about 100
mg CaCO3 L-1 or 2 meq L-1), then the pH may drop below
pH 7.0 (Ekama and Wenzel, 2008).
In the second stage of the nitrification process, also
known as the nitratation process, nitrite is oxidized to
nitrate by XNOO by means of the enzyme nitrite
oxidoreductase (Nir):
NO-2 + 0.5O2 → NO-3
Eq. 2.4.5
XNOO are aerobic chemolithoautotrophic bacteria
using, for synthesis, inorganic carbon as a carbon source
(e.g. HCO3-) and ammonium as a nitrogen source. When
including bacterial growth in the above equation, the
following metabolic macro chemical reaction equation is
obtained:
NO-2 + 0.003NH+4 + 0.485O2 + 0.015HCO-3 + 0.012H+ →
NO-3 + 0.009H2 O + 0.003C5 H7 O2 N
Eq. 2.4.6
The combination of the nitritation process performed
by XAOO and the nitratation process performed by XNOO
constitutes the nitrification process. The nitrification
stoichiometry can also be expressed according to the
standard ASM (activated sludge model) as:
1
4.75 ‒ YANO
+ iN,ANO · SNHx +
· SO2 +
YANO
YANO
Eq. 2.4.7
iN,ANO
2
1
+
· SIC → XANO +
· SNO3
14
14·YANO
YANO
Where, SNO3 is the nitrate concentration (mg N L-1);
SIC is the alkalinity concentration (mmol L-1), SO2 is the
DO concentration (mg O2 L-1), SNHx is the ammonium
concentration (mg N L-1), XANO is the concentration of
nitrifying organisms (mg COD L-1), YANO is the growth
yield of nitrifying microorganisms (g COD g N-1), and
ACTIVATED SLUDGE ACTIVITY TESTS
75
iN,ANO is the nitrogen content in the nitrifying organism’s
cells (g N g COD-1).
2.4.1.2 Denitrification
In the 1850s, first Reiset (1856) and then Pasteur (1859)
reported that the reduction of nitrate was of a biological
nature, marking the beginning of research into the
biological nitrogen cycle. Even if Pasteur erroneously
attributed nitrate reduction to ‘lactic yeast’, he
understood the role of organics. Reiset (1856) observed
that nitrogen is released into the atmosphere during the
decay of plant and animal residues. Shortly after, Gayon
and Dupetit (1883) named the process denitrification,
organics were experimentally proven to be required in the
process (Munro, 1886), and nitrite, nitric oxide (NO) and
nitrous oxide (N2O) were identified as intermediates
(Payne, 1986). It was observed that not only bacteria but
also eukaryotes and archaea can grow on the energy
gained by the oxidation of organics or inorganic
substrates coupled with the reduction of nitrate to nitrite,
NO, N2O and finally dinitrogen gas (N2) (RisgaardPetersen et al., 2006; Pina-Choa et al., 2010). Certain
microorganisms capable of performing heterotrophic
nitrification were also shown to be able to carry out
denitrification under aerobic and anoxic conditions, the
so-called aerobic denitrification (Robertson et al., 1995).
Denitrification may also proceed without N2O as an
intermediate as in the recently discovered denitrification
using methane as the electron donor (Ettwig et al., 2010).
Nitrifier denitrification where XAOO reduce nitrite to N2O
has also been reported (Bock et al., 1995). In contrast to
other microorganisms in the nitrogen cycle (e.g. XAOO,
XNOO, anammox), several ordinary heterotrophic
organisms (XOHO) are facultative denitrifiers that
preferentially use oxygen as an electron acceptor due to
the higher energy yield and only switch to denitrification
when low oxygen levels prevail in the presence of nitrate
or nitrite (Zumft, 1997). Independent of the electron
donor used, the overall biochemical pathway for
denitrification involves the same enzymes in each of
these reduction steps from nitrate to dinitrogen gas:
nitrate reductase (NAR), nitrite reductase (NIR), nitric
oxide reductase (NOR) and nitrous oxide reductase
(NOS):
NO-3 aq)
NAR
NO-2 aq)
NIR
NO(g)
NOR
N2 O(g)
NOS
N2 (g)
Eq. 2.4.8
If either the denitrifying microorganisms do not
express all the enzymes for the complete denitrification
chain or alternatively under certain environmental
conditions, then the intermediates NO and N2O can be
emitted, both of which have a negative impact on the
environment due to their toxicity and direct or indirect
contribution to the greenhouse effect.
In wastewater treatment systems, biological nitrogen
removal is mostly carried out by XOHO using organic
matter as the electron donor. When the organics naturally
present in the wastewater are not sufficient to achieve
complete denitrification, usually external electron donors
(such as acetic acid or methanol, among others) are
dosed. Due to its prevalence in wastewater applications,
only denitrification catalysed by XOHO (heterotrophic
denitrification) is considered in this chapter, which
occurs by the following generic catabolic pathway
(Mateju et al., 1992):
NO-3 + 1.25CH2 O → 0.5N2 + OH- + 0.75H2 + 1.25CO2
Eq. 2.4.9
This equation illustrates that heterotrophic
denitrification renders the environment more alkaline due
to the production of hydroxide ions.
XOHO can use both nitrate and nitrite as an electron
acceptor. The process of nitrate reduction to dinitrogen
gas is called denitrification, while the process of nitrite
reduction to dinitrogen gas is called denitritation.
According to standard activated sludge model (ASM)
notation, when a generic soluble and biodegradable
substrate (SB) is used as the carbon source, the
denitrification and denitritation stoichiometry can be
described as follows, respectively:
1 ‒ YOHO,Ax
1
· SB +
· SNO3 + iN,OHO · SNHx →
2.86 · YOHO,Ax
YOHO,Ax
1 ‒ YOHO,Ax
iN,OHO
‒
· SIC
XOHO +
2.86 · 14 · YOHO,Ax
14
1 ‒ YOHO,Ax
Eq. 2.4.10
+
· SN2
2.86 · YOHO,Ax
1 ‒ YOHO,Ax
·S
+i
·S
→
1.71 · YOHO,Ax NO2 N,OHO NHx
YOHO,Ax
1 ‒ YOHO,Ax
iN,OHO
XOHO +
‒
· SIC
1.71 · 14 · YOHO,Ax
14
1 ‒ YOHO,Ax
+
· SN2
Eq. 2.4.11
1.71 · YOHO,Ax
1
· SB +
Where, YOHO,Ax is the heterotrophic growth yield
under anoxic conditions (g COD g COD-1), XOHO is the
concentration of heterotrophic microorganisms (mg
COD L-1), SNHx is the ammonium concentration (mg N
L-1), SNO2 the nitrite concentration (mg N L-1), SNO3 the
76
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
nitrate concentration (mg N L-1), SN2 is the concentration
of dissolved dinitrogen gas (mg N L-1), and SIC is the
alkalinity concentration (mmol L-1).
2.4.1.3 Anaerobic ammonium oxidation (anammox)
The anammox process can be regarded as a peculiar type
of denitrification, in which the oxidation of ammonium
is coupled to the reduction of nitrite. Its discovery in the
1990s radically changed the understanding of the
biological nitrogen cycle (Kuypers et al., 2005), refuting
the conventional assumption at that time that ammonium
was chemically inert, and that its oxidation required
oxygen and a mixed-function oxygenase enzyme (van de
Graaf et al., 1996). Anammox bacteria belong to the
genus Planctomycetes and have been detected in several
wastewater treatment plants and natural environments all
around the world, showing their ubiquitous distribution.
Ammonium, nitrite and bicarbonate are the main
substrates in the anammox process (van de Graaf et al.,
1996). The catabolic reaction catalysed by anammox
bacteria couples the nitrogen atoms from ammonium and
nitrite to form dinitrogen gas (N2):
NH+4 + NO-2 → N2 + 2H2 O
Eq. 2.4.12
In the absence of oxygen, anammox bacteria activate
the stable ammonium molecule through the oxidizing
power of nitric oxide (NO). Briefly, anaerobic
ammonium oxidation is a three-step process with NO and
hydrazine as intermediates: first, nitrite is reduced to NO
by the enzyme nitrite oxidoreductase (Nir), then the
produced NO reacts with ammonium to form hydrazine
(N2H4), catalysed by the unique hydrazine synthase
enzyme (HZS), and finally hydrazine is oxidized to N2 by
hydrazine dehydrogenase (HDH) (Kartal et al., 2011).
Anammox bacteria are autotrophs and thus make use of
inorganic carbon as the carbon source for the production
of biomass. In the anabolic reaction the reducing
equivalents for the reduction of inorganic carbon
originate from the oxidation of nitrite to nitrate as
illustrated in the equation below (ammonium is
considered as the N source):
HCO-3 + 2.1NO-2 + 0.2NH+4 + 0.8H+ →
CH1.8 O0.5 N0.2 + 2.1NO-3 + 0.4H2 O
Eq. 2.4.13
The metabolic macro chemical reaction equation of
the anammox process is still subject to debate. Due to the
difficulties in cultivating a pure culture of anammox
bacteria, stoichiometric equations derived by
substrates/products via mass balance are intrinsically not
completely precise (Lotti et al., 2014). However, the first
reaction stoichiometry reported by Strous and co-authors
(1998) is widely used for reactor design and operation
purposes:
-
NH+4 + 1.32NO2 + 0.066HCO-3 + 0.13H+ →
1.02N2 + 0.066CH2 O0.5 N0.15 + 0.26NO-3 + 2.03H2 O
Eq. 2.4.14
Since nitrogen is usually present in wastewater as
ammonium but the anammox metabolism requires both
ammonium and nitrite as substrates, the anammox
process has to be combined with another process to
generate the required nitrite. For this purpose, the partial
nitritation (PN) process is usually applied. Nitrogen
removal processes based on the combination of PN and
anammox process are currently part of the state of the art
with about 100 full-scale implementations worldwide
treating mostly the effluent from anaerobic sludge
digesters as well as a variety of ammonium-rich
municipal and industrial wastewaters: leather tanning,
food processing, and the semiconductor, fermentation,
yeast production, distilling, and winemaking industries
(Lackner et al., 2014). Furthermore, encouraging results
were reported for pilot-scale installations treating black
water digestate (de Graaff et al., 2011), digested manure
(Villegas et al., 2011), urine (Udert et al., 2008) and
pharmaceutical wastewaters (Tang et al., 2011).
Recently, positive results have also been obtained in the
treatment of aerobically pre-treated sewage, opening up
new perspectives for converting municipal WWTPs from
energy-depleting to energy-generating systems (Lotti et
al., 2015a).
2.4.2 Process-tracking alternatives
According to the stoichiometry of the processes involved
in nitrogen removal, as presented in Section 2.4.1,
various alternatives are available to assess the process
kinetics and stoichiometric parameters of interest during
a batch test, such as:
• Chemical tracking by assessing the nitrite, nitrate or
ammonium concentrations over time; the choice of
the optimal chemical species to be tracked depends
on the specific process of interest.
• Titrimetric tracking by applying pH-static titration to
those processes that relevantly affect the solution pH.
• Manometric tracking, applicable to processes
involving soluble gaseous species of low solubility
such as N2.
• Respirometry, applicable to aerobic processes that
affect the DO concentrations.
ACTIVATED SLUDGE ACTIVITY TESTS
The last alternative is presented not in this chapter,
but in Chapter 3 on Respirometry, which is fully
dedicated to this technique.
The other tracking alternatives are briefly presented
hereafter.
2.4.2.1 Chemical tracking
Tracking the concentration over time of substrates and
products is the most common way to assess the kinetics
of a process.
When nitrification is to be tracked, ammonium and
nitrate concentrations are monitored over time. For the
separate assessment of the nitritation or nitratation rates,
ammonium and nitrite or nitrite and nitrate
concentrations have to be measured over time. Finally,
when nitritation and nitratation rates have to be assessed
simultaneously, ammonium, nitrite and nitrate
concentrations need to be monitored over time.
For denitrification, the most common batch test that
applies the chemical tracking procedure is the so-called
nitrate uptake rate (NUR) test. NUR tests make it
possible to assess several parameters of practical interest,
such as nitrate utilization rates, the utilization of organics
for denitrification and the anoxic biomass yield
coefficient (Naidoo et al., 1998; Kujawa and Klapwijk,
1999). As described in Section 2.2.4, samples are taken
during the course of the batch tests and then chemical
analyses of the main substrates (e.g. nitrite, nitrate and
COD) are performed in order to get sufficient
information for the assessment of relevant kinetic and
stoichiometric parameters.
Finally, when this method is applied to the evaluation
of the anammox process kinetics, ammonium, nitrite and
nitrate concentrations are monitored over time. Besides
the anammox process rate, this method also makes it
possible to assess stoichiometric parameters of interest
such as the ratio between nitrite and ammonium
consumption (NO2/NH4 ratio) and the ratio between
nitrate production and ammonium consumption
(NO3/NH4 ratio) which can be calculated as the ratio of
the corresponding conversion rates (Lotti et al., 2014).
2.4.2.2 Titrimetric tracking
The pH-static titration technique consists of the
controlled addition of an appropriately diluted solution of
acid or base to maintain a constant pH (therefore ‘static’
77
pH) in a biological system where the pH is affected by
different reactions. Under these conditions, the titration
rate is proportional to the reaction rate via a
stoichiometric factor. In principle, this technique is
applicable to any biological or physical-chemical
reaction affecting the proton concentration (thus, linked
with pH), i.e. any reaction converting neutral substrates
into acid or basic products, or acid or basic substrates into
neutral products. There are several biological reactions of
environmental interest that affect the pH of the
suspension where they take place. Attempts to use pHstatic titration have been mainly focused on nitrification
(Gernaey et al., 1997, 1998; Massone et al., 1998) and
denitrification (Massone et al., 1996; Rozzi et al., 1997;
Bogaert et al., 1997; Foxon et al., 2002). Although these
are consolidated applications of the pH-static titration
technique, others can be foreseen, such as those involving
the CO2/HCO3-/CO32- equilibria, e.g. the heterotrophic
degradation of organic substrates that produces CO2
(Ficara and Rozzi, 2004) and acetoclastic
methanogenesis that produces bicarbonate (Rozzi et al.,
2002).
Nitrification monitoring is the first and most
consolidated application of this technique since the
relationship between the ammonium oxidised and the
proton produced YNH4_H+ can be calculated from the
reaction stoichiometry, as assumed by ASM1 (Henze et
al., 2000):
YNH_H+ =
14
≈ 6.92 g N mol (protons)-1
2 + iN,ANO ·YANO
Eq. 2.4.15
The N/H ratio is therefore the stoichiometric factor
that allows the conversion of the titration rate into the
ammonium consumption rate. The pH-static titration
technique can also be applied to denitrification since this
process is a ‘pH-affecting reaction’. However, the
assessment of the ratio between nitrite or nitrate
consumption and proton production YNOx_H+ based on
stoichiometry is not as straightforward as for
nitrification, mostly because it depends on many more
factors, such as the carbon source, the sludge
characteristics, and the pH set point. To theoretically
calculate (YNOx_H+, a conceptual model was proposed by
Petersen et al. (2002), which considers that the following
four processes have a pH-dependent effect on proton
production during denitrification: (i) uptake of weak
organic acids as carbon source, (ii) uptake of nitrate, (iii)
uptake of ammonia for cell synthesis, and (iv) production
of carbon dioxide from organic carbon oxidation. Based
78
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
on these assumptions, the following reaction
stoichiometry was proposed to assess the ratio between
the net proton production and nitrate consumption:
1 ‒ YOHO,Ax
·S
+i
·S
→
β · YOHO,Ax NOx N,OHO NHx
YOHO,Ax
1 ‒ YOHO,Ax
a
·S + ‒
XOHO +
β · YOHO,Ax N2
C · YOHO,Ax
1 ‒ YOHO,Ax
b · iN,OHO C · 1 ‒ YOHO,Ax · x
‒
+
+
· H+
β · 14 · YOHO,Ax
14
C · YOHO,Ax
1
· SB +
Eq. 2.4.16
Where, SNOX is the nitrite or nitrate concentration; x
is the number of carbon moles per mole of organic
substrate, C is a factor (in g COD mol-1 organic substrate)
to express the organic carbon in COD units, β is the
oxygen equivalent of oxidized nitrogen; and a, b, c are
pH-dependent factors which take into account the
dissociation equilibria of weak acids/bases (a for organic
acids - HA, b for carbonic acid, and c for ammonium):
a=
b=
c=
A10-pKa
=
HA + A10-pH + 10-pKa
10pH - pK1 · (1 + 2 · 10pH - pK2 )
1 + 10pH - pK1 · (1+10pH - pK2 )
NH+4
NH+4 + NH3
=
10-pH
10-pH + 10-pKNH4
Eq. 2.4.17
Eq. 2.4.18
Eq. 2.4.19
Where, pKa is the dissociation constant for acetic acid
(4.75 at 25 °C), pK1 the dissociation constant for carbonic
acid (6.352 at 25 °C), pK2 the dissociation constant for
bicarbonate (10.33 at 25 °C), and pKNH4 the dissociation
constant for ammonium (9.25 at 25 °C).
By substituting the correct value of β (i.e. 2.86 g COD
N-1 for N-NO3 and 1.72 g COD N-1 for N-NO2), it follows
that YNOx_H+, in g N mol-1, in the presence of nitrate,
YNO3_H+, or nitrite, YNO2_H+, as an electron acceptor, can
be expressed as:
1 ‒ YOHO,Ax
YNO3_H+ =
·
2.86 · YOHO,Ax
1 ‒ YOHO,Ax
c · iN,OHO
a
‒
‒
+
+
2.86 · 14 · YOHO,Ax
14
C · YOHO,Ax
b · 1 ‒ YOHO,Ax · x
C · YOHO,Ax
-1
Eq. 2.4.20
1 ‒ YOHO,Ax
·
1.72 · YOHO,Ax
1 ‒ YOHO,Ax
c · iN,OHO
a
‒
+
+
‒
14
C · YOHO,Ax 1.72 · 14 · YOHO,Ax
YNO2_H+=
b · 1 ‒ YOHO,Ax · x
C · YOHO,Ax
-1
Eq. 2.4.21
These equations show that the assessment of YNOx_H+
makes it necessary to know the chemical composition of
the carbon source (C and x), which is seldom the case in
practical applications, and the anoxic biomass growth
yield coefficient, YOHO,Ax. As such, its evaluation is
theoretically possible, but difficult in practice.
Fortunately, YNOx_H+ can also be experimentally
evaluated by measuring the amount of titration solution
(normally acid) dosed under pH-static conditions to
denitrify a known amount of nitrite or nitrate and in the
presence of the carbon source of interest. Once the
YNOx_H+ ratio is measured, the titration rate can be easily
converted into the nitrate or nitrite uptake rate.
Theoretically, even the anammox process can be
monitored by pH-static titration. However, there is very
little available experience of this, and therefore this
alternative is not discussed in this chapter.
2.4.2.3 Manometric tracking
According to this technique, the rate of a bioprocess that
produces a poorly soluble gaseous component is
proportional to the rate of increase in pressure, provided
that the bio-reaction takes place in a gas-tight reactor.
The relationship between the generated overpressure,
P(t), and the volumetric gas production, VG(t), can be
obtained by assuming that the gas transfer from the liquid
to the gas phase is not rate-limiting (sludge mixing allows
the quick transfer of gaseous species) and that no relevant
amounts of the gaseous species remain in solution. Under
these conditions and at constant temperature, according
to the gas law, the following relationship applies:
VG (t) =
P(t ) ‒ Patm
· VHS
Patm
Eq. 2.4.22
Where, VHS is the volume of the headspace in the
reactor and Patm is the atmospheric pressure.
This measuring principle was proven to be applicable
and advantageous in the monitoring of denitrification
(Sánchez et al., 2000; Ficara et al., 2009) and of
anaerobic ammonia oxidation (Dapena Mora et al., 2007;
ACTIVATED SLUDGE ACTIVITY TESTS
Scaglione et al., 2009; Bettazzi et al., 2010; Lotti et al.,
2012) since both processes produce dinitrogen gas. As
for denitrification, a CO2 adsorbent should be used that is
typically located in the gas headspace (e.g. NaOH pellets)
so that overpressure data are only related to the release of
N2.
79
•
•
2.4.3 Experimental setup
2.4.3.1 Reactors
Independently of the technology applied to remove
nitrogen from wastewater, to assess the performance of
the biological nitrogen removal process, batch activity
tests can be carried out under aerobic (nitrification) or
anoxic conditions (denitrification and anammox)
depending upon the parameters of interest and nature of
the study. In any case, the reactor(s) used for the
execution of tests must have the required means to: (i)
avoid oxygen intrusion under anoxic conditions, (ii)
secure satisfactory SO2 availability under aerobic
conditions (e.g. SO2 higher than 2 mg L-1), (iii) provide
satisfactory mixing conditions, (iv) maintain an adequate
and desirable temperature, (v) provide precise pH
control, and (vi) have additional ports for sample
collection and addition of influent, solutions, gases and
any other liquid media or substrate used in the test. For
the requirements needed to ensure proper anoxic
conditions, aerobic conditions, mixing, temperature
control, pH control and sampling and dosing ports during
the execution of the tests, the reader is referred to Section
2.2.2.1. However, when titrimetric or manometric
experiments are to be performed, special apparatus
should be available, as described below.
•
operation is required when performing denitrification
tests.
Probes to assess: temperature (with a resolution of 0.1
°C), pH (with a resolution of 0.01 pH units) and,
possibly, DO (± 0.02 mg L-1 around the selected set
point value).
An aeration system (for aerobic processes), typically
with an aeration capacity of 50-200 L L-1 h-1, and
stone diffuser for fine bubble aeration;
An automated titration system capable of maintaining
the pH within a narrow range (e.g. at a defined pH set
point ± 0.02) and to record the volume of the titration
solution dosed over time with a suggested resolution
of 0.1 mL and a minimum logging frequency of 1
datum per minute. Manual recording of the added
titration solution volume can also be obtained by
storing the titration solution in a graduated cylinder
and by manually reading the remaining volume at
regular intervals (every few minutes for nitrification
tests) or by installing the solution on a balance with
its weight either read or recorded automatically. A
solution of 0.05-0.02 N NaOH can be used as an
alkaline titration solution, while a 0.05-0.02 N HCl
solution can serve as an acidic titration solution.
9
10
7
2.4.3.2 Instrumentation for titrimetric tests
To perform set point titration tests, an automated titration
unit is required. As for respirometers, such systems are
readily available on the market. However, they can be
easily implemented by using conventional laboratory
equipment and basic signal acquisition and control units.
Specifically, an automated titration unit should be made
up of the following components (see Figure 2.4.1):
• A reaction vessel: a well-mixed reactor with a
thermostat or temperature control to host the
activated sludge sample with an operative volume of
0.5 to 1.5 L. The reaction vessel does not need to be
gas-tight when performing nitrification tests.
However, a reduced solid-liquid contact area is
preferable to limit gas-liquid transfer of oxygen and
carbon dioxide. Gas tightness for proper anoxic
8
3
4
5
1
6
2
Figure 2.4.1 Scheme of a pH-static titration system: 1. aerator; 2. mixer;
3. temperature probe; 4. pH probe; 5. DO probe; 6. reaction vessel; 7.
titration-solution dosing system; 8. signal acquisition and recording; 9.
alkaline titration solution; 10. acidic titration solution.
80
When aerobic bioprocesses are involved, the pHstatic system can be conveniently upgraded into a
pH/DO-stat system, in which a secondary titration unit
provides an H2O2 diluted solution, serving as an
oxygenated titration solution. For this purpose, the
system described in Figure 2.4.1 should be upgraded by
integrating:
• A DO probe, with a minimum resolution of 0.1 mg
L-1.
• An additional automated dosing or titration system
capable of maintaining the SO2 within a narrow range
(± 0.1 mg L-1 around the selected set point value). A
0.05-0.2 M H2O2 solution is appropriate when
performing pH/DO-static titration tests on
conventional activated sludge samples. When dosed,
the H2O2 solution will be converted into molecular
oxygen (O2) and water by peroxydases produced by
aerobic bacteria to counteract oxidative stress, and
thus making oxygen available for bacterial
respiration. As a matter of fact, it has been observed
that diluted H2O2 solutions can be used for short-term
respirometric tests without significant bacteria
inhibition (e.g. Ficara et al., 2000).
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
•
•
d. A manometric measuring device, possibly
featured with a data logger, and fixed on the top
of a glass bottle with a resolution of 1-3 mbar.
A constant temperature incubator that limits
temperature oscillations to ± 0.2 °C. It is mandatory
that temperature is very well controlled during the
course of the test since temperature variations cause
changes in the overpressure values that are not
associated with gas release and this would therefore
result in data noise.
A magnetic stirrer that can operate at around 100-200
rpm. Alternatively, a thermostatic orbital shaker can
be used to ensure both temperature control and
mixing. In the case of anammox, a magnetic stirrer is
advisable only in the case of suspended anammox
biomass (100-200 rpm), while for both hybrid,
biofilm on carriers and granular anammox biomass
types (Hu et al., 2013), a shaker is preferred in order
to avoid deterioration of the anammox biofilm due to
the shear forces caused by the magnetic stirrer.
The scope of this DO-static titration unit is to
maintain the DO value at a predefined set point level
(DO-set point) by titrating the oxygenated titration
solution, thus meeting the following objectives:
• To maintain the desired redox condition without the
need for air bubbling; this makes it possible to avoid
CO2 stripping which is a pH-affecting process that
overlaps with other targeted pH-affecting reactions.
• To assess the oxygen consumption rate of the reaction
that, under DO-static operation, equals the titration
rate of the oxygenated titration solution. This is
additional information that can be used to check or
complement the titration rate of the alkaline/acidic
solution, as described in detail later on in this chapter.
2.4.3.3 Instrumentation for manometric tests
Tests should be performed by using gas-tight apparatus.
Typically, systems applied to perform BOD tests are
used. The minimum requirements are the following:
• A glass bottle with (see Figure 2.4.2):
a. A working volume of around 1 L.
b. Two lateral openings, sealed by rubber septa kept
in place by plastic or aluminium gear, for
substrate injections and gas flushing/discharge.
c. A container for NaOH pellets located in the bottle
headspace and serving as a CO2 trap.
Figure 2.4.2 Commercially available apparatus to perform manometric
tests for denitrification purposes (photo: Lotti, 2016).
ACTIVATED SLUDGE ACTIVITY TESTS
2.4.3.4 Activated sludge sample collection
The sampling time and location of an activated sludge
sample performing nitrification and denitrification is
highly dependent on the type of batch activity test to be
conducted. The removal of nitrogen operated via
nitrification and denitrification processes is based on the
alternating aerobic-anoxic conditions. Thus, preferably, a
fresh sample should be collected at the end of the relative
reaction stage: aerobic for nitrification, anoxic for
denitrification. Certainly, the sampling location will
depend on the system configuration. For instance, the
sampling time and location of an activated sludge sample
from a PN/anammox system depends on the type of
technology used; obviously when PN and anammox
processes are divided into two separate stages, anammox
biomass should be sampled by the anammox stage. In
full- and pilot-scale wastewater treatment plants, the
physical boundaries between stages must be identified
prior to sampling. In extreme cases, where the phases are
not (physically) well defined, the redox limits or
boundaries need to be determined with the use of a DO
meter, redox meter and/or by determination of the nitrate
and nitrite concentrations. In lab-scale systems (usually
operated on a time-base mode), the sample collection can
be relatively easier, since the reaction time defines the
length of the stages. To obtain homogenous and
representative samples, the sludge samples must be
collected in sampling spots where well-mixed conditions
take place.
When anammox is the targeted process, then
sampling from the outlet of the anammox tank is
typically adequate. This sampling point ensures that the
sludge samples collected contain limited amounts of
residual ammonium/nitrite concentrations. When
granular anammox biomass is considered, the outlet of
the anammox tank normally contains very few granules
because a granular system is usually equipped with a
biomass retention system (e.g. a three-phase separator,
hydro-cyclone, settling phase in a sequencing batch
reactor (SBR) cycle, etc.). In the case of anammox
granular systems then sampling the mixed liquor directly
from the anammox tank may be adequate in continuously
operated systems (CSTR, continuous stirred tank
reactor), while for SBR systems, sampling should be
performed before the settling phase when the reactor is
completely mixed to ensure completely mixed conditions
and the presence of limited amounts of residual
ammonium/nitrite concentrations.
81
Ideally, batch activity tests must be performed as
soon as possible after sample collection. In lab-scale
systems, in principle, this should not be a problem if the
batch activity tests are performed in the same laboratory
and their execution is coordinated and synchronized with
the operations of the lab reactor. Also, at full- and pilotscale treatment plants, batch activity tests can be
performed in situ shortly after mixed liquor collection if
the sewage plant laboratory is conditioned and equipped
with the required experimental and analytical equipment.
If the batch activity tests cannot be performed in situ on
the same day, a mixed liquor sample can be collected and
transferred to the location where the tests will be
executed. Thereafter, the sampling bucket can be
properly stored and transported in a fridge or in ice box
(below or close to 4 °C) under non-aerated conditions and
the activity tests should be performed no later than 24 h
after sampling. In order to avoid the creation of anaerobic
conditions during storage and the undesired production
of toxic sulphide through the reduction of sulphate,
nitrate should be added to the mixed liquor at a final
concentration of about 50-200 mg N L-1. However, the
availability of nitrate would promote endogenous
biomass respiration. Therefore, it is important to stress
that it is highly recommended to conduct the test as soon
as possible after sample collection.
Especially after storage, the biomass present in the
mixed liquor sample needs to be ‘washed’ to remove any
added nitrate, ‘re-activated’ and acclimatized to the target
pH and temperature of interest prior to the execution of
the batch activity tests. The washing step must be
performed by using an appropriate ‘washing medium’
with a mineral composition that depends on the target
microbial population, as indicated below. Tap water can
also be used as a washing medium as long as its
conductivity is similar to that of the cultivation medium.
In any case, the in situ execution of the batch activity tests
is preferable since this avoids the exposure of biomass to
varying conditions. The total volume of activated sludge
(mixed liquor) to be collected depends on the number of
tests, reactor volume and total volume of samples to be
collected to assess the biomass activity. Often, 10-20 L
of activated sludge or mixed liquor from full-scale
wastewater treatment plants is considered sufficient. On
the other hand, samples collected from lab-scale reactors
rarely reach more than 1 L because lab-scale systems are
usually smaller (from 0.5 to 2.2 L and in certain cases up
to 8-10 L) and the maximum volume that can be
withdrawn from lab-scale reactors is often set by the
daily withdrawal of the excess of sludge from the system.
Since the maximum volume allowed to be withdrawn is
82
directly related to the applied solids retention time (SRT),
which is defined by the growth rate of the organisms,
particular attention must be paid when dealing with slowgrowing organisms such as nitrifiers and anammox
bacteria.
Suggestions on sampling scheduling and ideal
storage times have been previously described (see
Section 2.2.3) and should be carefully considered.
2.4.3.5 Activated sludge sample preparation
Generally speaking, activated sludge samples can be used
as such or after specific adjustments in pH (with or
without the presence of a pH buffer), temperature,
ammonium/nitrate/nitrite concentration, carbon source
concentration, and XVSS. For conventional activated
sludge samples from wastewater treatment plants treating
urban wastewater, a sample which has a XVSS around 2-4
g VSS L-1 would be ideal. For anammox sludge, XVSS
around 5-10 g VSS L-1 will be preferable. For very
diluted or concentrated sludge samples, a preconcentration step (e.g. by decanting into an Imhoff cone
for 30 min or by centrifuging at 4,000 rpm for a few
minutes) or dilution with the secondary effluent of the
same wastewater treatment plant may be helpful. This
will avoid the occurrence of too slow or too fast
conversion rates.
When such procedures are implemented on anoxic or
anaerobic activated sludge samples, N2 sparging should
be performed immediately afterwards in order to reestablish proper anoxic/anaerobic conditions. When the
anammox process is considered, a mixture of N2/CO2
gases (usually 95/5 % is used in practice) can be used for
sparging instead of N2 in order to avoid excessive CO2
stripping which would cause a pH increase and may limit
anammox activity during the batch test due to limiting
inorganic carbon concentration. As for pH and
temperature, in principle, the closer the set point pH
value is to the typical operational pH of the plant, the
more the resulting process rate will be representative of
the operational process rate. The same concept applies to
the selection of the temperature value. Since the effect of
the anammox process on the pH is rather limited (0.13
mole of protons consumed per mole of ammonium
converted), limited variations on the pH are expected
during the execution of a batch test (e.g. from 7.5 to 7.9
according to Lotti et al., 2012). Nevertheless a pH buffer
such as Hepes (N-2-hydroxyethyl-piperazine-N0-2ethane sulfonic acid) or phosphate can be used to
maintain a constant pH throughout the duration of the
batch test (Dapena-Mora et al., 2007; Lotti et al., 2012).
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
While a Hepes buffer can be used at concentrations up to
25 mM without affecting anammox activity (Lotti et al.,
2012), the concentration of the phosphate buffer should
be carefully decided since it may result in process
inhibition (Dapena-Mora et al., 2007; Oshiki et al.,
2011). Previous reports have shown that a phosphate
buffer concentration of 5.3 mM is suitable for the
conduction of anammox batch tests (Dapena-Mora et al.,
2007; Lotti et al., 2012).
In general, the objective of the sampling and
experimental campaigns should be to minimize as much
as possible the need for transportation, cooling, storage
and reactivation of the sludge. Whenever possible, it is
advisable to use 'fresh' sludge (and substrate/media).
When actual operational conditions are to be tested, then
the corresponding batch activity tests must be executed
right away after sludge collection with the minimum
adjustments of the operational conditions (e.g. for pH and
temperature). When the batch tests cannot be performed
in situ or shortly after collection, sludge samples must be
stored at around 4 °C for preservation purposes during
transportation and storage. In the case of anammox
biomass, ambient temperature can be adopted for
preservation purposes during transportation and storage.
Storage at ambient temperature is suggested to avoid
temperature shocks from the usual operative temperature
of an anammox system (25-35 °C) to 4 °C, which is
usually considered an adequate storage temperature for
conventional activated sludge samples. Under these
circumstances, the batch activity tests should preferably
be executed in less than 24 h after sludge collection and
after 'reactivation' by keeping the sludge at the pH and
temperature of interest (after N2 flushing in case of
denitrification and anammox).
The addition of limited amounts of the substrate can
favour bacterial metabolic reactivation. However, the
preparation of the activated sludge is test-dependent and
specific suggestions/recommendations are described in
the following paragraphs.
2.4.3.6 Substrate
When real wastewater (either raw or settled) is used for
the execution of activity tests, it can be fed in a relatively
straightforward manner to the reactor/fermenter. A rough
filtration step (using 10 μm pore filter size) can be used
to remove the remaining debris and large particles
present in the raw wastewater.
If different carbon sources and concentrations are to
be studied, the plant effluent can also be used to prepare
ACTIVATED SLUDGE ACTIVITY TESTS
a semi-synthetic media containing a SB concentration of
between 50 and 100 mg COD L-1.
For the execution of conventional denitrification
tests, nitrate and nitrite solutions can be prepared to
create the required anoxic conditions. For this purpose,
different stock solutions can be prepared using nitrateand nitrite-salts.
For the execution of anammox tests, ammonium and
nitrite need to be dosed to the wastewater at the beginning
of the test in order to provide the desired amount of
substrate. For this purpose, different stock solutions can
be prepared using ammonium- and nitrite-salts (e.g. 1-10
g N L-1); the most commonly used are ammonium
sulphate and sodium nitrite, respectively. Particular
attention needs to be paid when considering the initial
substrate concentrations. Nitrite in fact, besides being a
substrate for the anammox process, is also an inhibitor
(Lotti et al., 2012; Puyol et al., 2014). Initial nitrite
concentration around 50-70 mg N L-1 is usually
considered adequate. Nevertheless, lower initial nitrite
concentrations (e.g. 10 mg N L-1) are recommended if the
anammox biomass originates from systems operated at
very low (few mg N L-1) nitrite concentrations (e.g.
DEMON systems, Wett et al., 2007). In fact, the nitrite
inhibition effect and resilience seems to depend on the
‘cultivation history’ of the anammox biomass, being the
cultures cultivated under strict nitrite limitation more
prone to nitrite inhibition (Lotti et al., 2012).
When tests are executed to assess the potential
inhibitory or toxic effect of a given compound at different
concentrations, concentrated stock solutions can be
prepared and added during the test to obtain the
concentrations of interest. Tests performed to assess
whether the inhibitory or toxic effects are reversible must
be carried out after washing the biomass to remove the
inhibiting or toxic compound(s). Often, the washing step
is performed by consecutive settling and re-suspension of
the sludge sample in carbon-free and nitrite-free media
(either fully synthetic or using a treated effluent after
filtration) under anoxic conditions.
2.4.3.7 Analytical procedures
Analytical procedures of interest (NH4, NO2, NO3,
MLSS, MLVSS, COD, BOD) should be performed
following standardized and commonly applied analytical
protocols detailed in Standard Methods (APHA et al.,
2012).
83
If a specific carbon source is used, its determination
should be performed according to the relevant analytical
method. However, most of the time, the determination of
soluble COD can be appropriate to follow the carbon
source utilization.
2.4.3.8 Parameters of interest
Nitrification
The most significant kinetic parameter of the aerobic
ammonium oxidation process (nitritation) performed by
AOO is the maximum biomass-specific ammonium
oxidation rate (qAOO,NH4). Similarly, for the aerobic nitrite
oxidation process (nitratation) performed by NOO, the
main parameter of interest is the maximum biomassspecific nitrite oxidation rate (qNOO,NO2_NO3). Table 2.4.1
presents typical kinetic parameter values found in
literature for both the aerobic ammonium and nitrite
oxidation processes. Literature values of the biomass
growth yield of both AOO and NOO are also reported in
Table 2.4.1.
Table 2.4.1 Expected stoichiometric and kinetic parameters of interest for
activated sludge performing aerobic ammonium and nitrite oxidation to
nitrite and nitrate. The kinetic parameters are reported considering a
reference temperature of 20 °C.
Aerobic ammonium oxidation–nitritation process
qAOO,NH4
YAOO
Reference
g N g VSS-1 d-1
g VSS g N-1
0.11
0.14
Blackburne et al. (2007)
0.09
0.11
Jones et al. (2007)
0.24
0.13
Jubany et al. (2008)
0.27
0.15
Koch et al. (2000)
0.21
0.11
Lochtman (1995)
0.22
0.15
Wiesmann (1994)
Aerobic nitrite oxidation–nitratation process
qNOO,NO2_NO3
YNOO
Reference
g N g VSS-1 d-1
g VSS g N-1
0.21
0.07
Blackburne et al. (2007)
0.13
0.07
Jones et al. (2007)
0.39
0.06
Jubany et al. (2008)
1.78
0.02
Koch et al. (2000)
0.45
0.03
Lochtman (1995)
0.78
0.04
Wiesmann (1994)
1.07
0.03
Wik and Breitholtz (1996)
84
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
The kinetics reported in Table 2.4.1 refer to an
operational temperature of 20 °C and were calculated
from the original values according to the Arrhenius
equation reported below.
kS T) = kS Tref ) · exp
Ea,S · TK ‒ Tref )
R · TK · Tref
Eq. 2.4.23
sludge system this can be directly related to the COD/N
ratio in the influent, with the lower COD/N ratio
corresponding to the higher fraction of active nitrifying
biomass. As depicted in Table 2.4.1, the reported growth
yield values are higher for the ammonium-oxidizing
bacteria than for the nitrite-oxidizing bacteria.
Denitrification
Where, kS(T) is the maximum biomass-specific
consumption rate of the substrate S evaluated at the
desired operative absolute temperature TK (K), Tref is the
reference absolute temperature, R is the ideal gas
constant (8.31 J mol-1 K-1), Ea,S is the activation energy of
the considered bioprocess consuming the substrate S.
Ea,NH4 = 68 kJ mol-NH4-1 K-1 and Ea,NO2 = 44 kJ mol-NO2-1
K-1 are typical activation energy values for the nitritation
process performed by AOO and the nitratation process
performed by NOO, respectively.
As it can be observed in Table 2.4.1, the aerobic
consumption rates for ammonium and nitrite oxidation
reported in literature can vary widely, especially for the
nitratation process catalysed by NOO. The main reason
can be the active biomass fraction present in the activated
sludge biomass, conventionally referred to as the total
MLVSS concentration. The larger the fraction of active
nitrifying biomass, the higher the specific biomass
conversion rate expected. In a conventional activated
To quantify the activity of OHOs under anoxic
conditions, the relevant stoichiometric parameters and
kinetic constants have to be known. As for stoichiometry,
the most relevant parameter to be defined is the
heterotrophic growth yield under anoxic conditions
YOHO,Ax (in g COD-biomass per g COD-substrate). Like
the aerobic growth yield, this parameter may depend on
various factors, such as the organic carbon source
quantity and quality, and the environmental conditions.
Typical values for this parameter are listed in Table 2.4.2.
This parameter can be easily determined by setting up
appropriately designed batch activity tests, as explained
later on. When performing denitrification, the carbon
source required per nitrate/nitrite to be removed
(expressed as the COD/N ratio or as the C/N ratio) can be
used instead of YOHO,Ax, given that a stoichiometric
relationship exists between the two of them. The COD/N
ratio represents the denitrification capacity of a carbon
source or of a wastewater and can be a more practical
parameter than the YOHO,Ax.
Table 2.4.2 Expected stoichiometric and kinetic parameters of interest for activated sludge wastewater treatment systems performing denitrification.
Parameter (symbol)
Heterotrophic anoxic growth yield (YOHO,Ax)
COD to nitrogen ratio (COD/N)
Maximum biomass-specific denitrification rate (qNOx_N2)
Remark
Acetate
Wastewater
Acetate
Ethanol
Methanol
Methanol
Methanol
Ethanol
Acetate
Acetate,
Acetate, nitrite
Methanol, 13 °C
Ethanol, 13 °C
Acetate, 13 °C
Acetate
Acetate
Value
0.66 g COD g COD-1
0.50 g COD g COD-1
0.66 g VSS g COD-1
0.22 g VSS g COD-1
0.18 g VSS g COD-1
4.6 g COD g N-1
4.7 g COD g N-1
3.5 g COD g N-1
3.4 g COD g N-1
10-19 mg N g VSS-1 h-1
15-28 mg N g VSS-1 h-1
9.2 mg N g VSS-1 h-1
30.4 mg N g VSS-1 h-1
31.7 mg N g VSS-1 h-1
1-3 mg N g VSS-1 h-1
2-10 mg N g VSS-1 h-1
Reference
Ficara and Canziani (2007)
Orhon et al. (1996)
Kujawa and Klapwijk (1999)
Hallin et al. (1996)
Tchobanoglous et al. (2003)
Bilanovic et al. (1999)
Mokhayeri et al. (2006)
Mokhayeri et al. (2006)
Mokhayeri et al. (2006)
Ficara and Canziani (2007)
Ficara and Canziani (2007)
Mokhayeri et al. (2008)
Mokhayeri et al. (2008)
Mokhayeri et al. (2008)
Kujawa and Klapwijk (1999)
Henze (1991)
ACTIVATED SLUDGE ACTIVITY TESTS
85
Regarding the growth kinetics, the specific substrate
consumption (either nitrate/nitrite or COD) can be easily
determined by batch activity tests. This value depends on
the operational conditions used during the test (especially
on temperature and the substrate’s nature and
concentrations). Therefore, these values should always
be specified when reporting the results of a test. As
reference conditions, the denitrification tests should be
performed at 20 °C under non-limiting concentrations of
carbon and nitrate/nitrite. This will allow the
determination of the maximum specific denitrification
rate. In practical applications, the maximum biomass
specific denitrification rate qNOx_N2 is linked to the mixed
liquor suspended solids (XTSS) or, more commonly, to
their volatile suspended solids content (XVSS). Values
reported in literature can be found in Table 2.4.2.
Anammox
The most significant kinetic and stoichiometric
parameters of the anaerobic nitrogen removal process
performed by anammox bacteria are the maximum
specific biomass ammonium oxidation rate (qAMX, NH4_N2),
the nitrite to ammonium (YNH4_NO2,AMX) consumption
ratio and the ratio between nitrate production and
ammonium consumption (YNH4_NO3,AMX). In Table 2.4.3
typical kinetic and stoichiometric parameters values
found in literature are reported, together with the biomass
growth yield.
Table 2.4.3 Stoichiometric and kinetic parameters of interest for anammox biomass performing the anaerobic ammonium oxidation process. The kinetic
parameters are obtained at 30 °C. Biomass type: suspended (S), flocculent (F), granular (G). Reactor of origin: lab- (Lab) or full-scale (Full) reactor
performing the anoxic stage of a 2-stage PN/anammox system (2-stage) or the 1-stage PN/anammox system (1-stage).
qAMX, NH4_N2
YAMX,NH4
YNH4_NO2,AMX
YNH4_NO3,AMX
g N2-N g VSS-1 d-1
0.66
0.22
0.22
2.01
3.38
0.16
0.55
C-mol NH4-mol-1
0.066
mol mol-1
1.32
1.27
1.28
1.22
mol mol-1
0.26
0.34
0.37
0.21
0.105
0.071
0.071
This parameter, even though it cannot be directly
measured through batch tests, is useful to convert the
specific biomass activity to growth rate values. Different
from other bioprocesses described in this section, the
anammox kinetics reported in Table 2.4.3 correspond to
values measured at 30 °C which is the most common
temperature at which the anammox process is operated in
both lab- and full-scale systems, since it is close to the
optimal temperature of these organisms (Hu et al., 2013).
As observed in Table 2.4.3, the anammox kinetic
rates reported in literature can vary widely. The main
reason for such variation appears to be the fraction of
active anammox biomass present in the sample. In
anammox reactors fed with autotrophic synthetic media,
as is often the case for a lab-scale system, a lower fraction
of XOHO is expected compared to reactors fed with CODcontaining wastewaters. The SRT applied is also
expected to affect the fraction of active cells, which may
be reduced by the accumulation of a significant fraction
of inactive cells and non-biodegradable matter at higher
Biomass type
F
G
F
S
S
G
G
Reactor of origin
Reference
Lab, 2-stage
Lab, 2-stage
Lab, 2-stage
Lab, 2-stage
Lab, 2-stage
Full, 1-stage
Full, 2-stage
Strous et al. (1998)
Puyol et al. (2013)
Puyol et al. (2013)
Lotti et al. (2014)
Lotti et al. (2015b)
Lotti et al. (2015c)
Lotti et al. (2015c)
SRT due to decay. Also, certain differences are observed
between the kinetics of 1- or 2-stage PN/anammox
systems because of the presence of XAOO in the former,
which contributes to a reduction in the fraction of active
anammox biomass. Finally, since the anammox process
is normally operated under nitrite limiting conditions in
view of the inhibition potential of this substrate (Lotti et
al., 2012), systems where biomass does not tend to
aggregate are characterized by lower mass transfer
limitations such as flocculent and (especially) suspended
sludge, having higher active anammox biomass fractions
compared to biofilm systems.
As observed in Table 2.4.3, also the
consumption/production stoichiometric ratios may vary
for different anammox systems. In literature, there is
evidence that the anammox stoichiometry can be affected
by the physiological state of the biomass, which can be
influenced by the N-load (Dosta et al., 2008; Yang et al.,
2009), temperature (Dosta et al., 2008) or pH (CarvajalArroyo et al., 2013).
86
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.4.3.9 Type of batch tests
A comprehensive list of tests that are described later
on in this chapter is presented in Table 2.4.4.
Depending on the type of process of interest
(nitrification, denitrification, anammox) and on the
selected tracking technique (chemical, titrimetric or
manometric), various tests can be performed to assess the
nitrogen removal conversions.
In the following paragraphs, these tests are described
in detail. First, nitrification tests are presented (Section
2.4.4), then denitrification tests (Section 2.4.5) and
finally anammox tests (Section 2.4.6).
Table 2.4.4 Batch activity tests performed to assess the biological nitrogen removal conversions as a function of the process and tracking method.
Test code
NIT.CHE
NIT.TIT.1
NIT.TIT.2
DEN.CHE.1
DEN.CHE.2
DEN.MAN
DEN.TIT
AMX.CHE
AMX.MAN
Process
Nitrification
Nitrification
Nitrification
Denitrification
Denitrification
Denitrification
Denitrification
Anammox
Anammox
Tracking method
Chemical
Titrimetric
Titrimetric
Chemical
Chemical
Manometric
Titrimetric
Chemical
Manometric
Purpose
Assessing the maximum NH4 oxidation rate
Assessing the maximum NH4 oxidation rate
Assessing the maximum NH4 and NO2 oxidation rate and the ammonification rate
Assessing the maximum denitrification rate and the anoxic growth yield on a specific C source
Assessing the denitrification potential of a wastewater
Assessing the maximum denitrification rate
Assessing the maximum denitrification rate
Assessing the maximum anammox rate and the NO2/ NH4 and NO3/ NH4 ratio
Assessing the maximum anammox rate
2.4.4 Nitrification batch activity tests:
preparation
2.4.4.1 Apparatus
Each tracking methodology (chemical, titrimetric or
manometric) has specific apparatus requirements. When
the chemical tracking is applied, refer to the following
list of equipment:
1. A batch reactor equipped with mixing system and
adequate sampling ports (Section 2.4.3.1).
2. A
calibrated
pH
electrode
(if
not
included/incorporated in the batch reactor setup).
3. A 2-way pH controller for HCl and NaOH addition
(alternatively a one-way control - generally for HCl
addition - or manual pH control can be applied
through the manual addition of HCl and NaOH). For
alkaline solutions that needs to be acidified, sparging
with gaseous CO2 (or a gas mixture enriched in CO2)
can be considered instead of the addition of an acidic
solution since it has the advantage of avoiding the
addition of the counter ions of protons (i.e. Cl- when
using HCl as the acidic solution).
4. A thermometer (with a recommended working
temperature range of 0 °C to 40 °C).
5. A temperature control system (if not included in the
batch reactor setup).
6. A
DO
meter
with
electrode
(if
not
included/incorporated in the batch reactor setup) to
verify the aerobic/anoxic conditions.
7. A stopwatch.
A list of the equipment required for titrimetric and
manometric tests is described in Section 2.4.3.1.
2.4.4.2 Materials
For general instructions on material preparation, refer to
Section 2.2.3.4 and Table 2.2.2. Test-specific
requirements will be listed within each protocol of the
test. For a complete list of the required materials refer to
Section 2.2.3.2 (with the exception of points 7 and 8).
2.4.4.3 Media preparation
• Real wastewater
For batch tests that require the use of a real wastewater,
follow the instructions reported in Section 2.2.3.3.
• Titration solutions
a. NaOH and HCl solutions are needed. Typically 0.050.1 N solutions would be suitable for most
applications.
b. The H2O2 solution can be obtained by dilution of
commercially available H2O2 solutions. The most
common H2O2 solution is 3 %, corresponding to a
ACTIVATED SLUDGE ACTIVITY TESTS
concentration of 0.44 mol O2 L-1. Therefore, an
appropriate oxygenated titration solution can be
obtained by diluting this solution 10 times (a final
concentration of 44 mmol O2 L-1). To check the
concentration of the H2O2 solution, the iodometric
method can be applied (method 4,500-Cl B in APHA
et al., 2012). The diluted solution should be stored in
dark bottles and new solutions should be prepared
every 7-10 days.
• Ammonium and nitrite stock solutions
These can be prepared from salts (e.g. from NH4Cl and
NaNO2). An adequate concentration of the stock
solutions is between 5 and 10 g N L-1. The pH of the
ammonium solution should be adjusted to 7.0 to reduce
any potential interference during the pH-static tests.
• Allyl-N-thiourea (ATU)
A stock solution of approximately 5-10 g L-1 is generally
adequate.
• Acid and base solutions
These should be 100-250 mL of 0.2 M HCl and 100-250
mL 0.2 M NaOH solutions for automatic or manual pH
control, and 10-50 mL of 1 M HCl and 10-50 mL 1 M
NaOH solutions for initial pH adjustment if the desired
operational pH is very different from the pKa value of the
buffering agent. Instead of NaOH, 0.2 M Na2CO3 can be
used as the alkaline solution, which has the advantage of
acting as both the base and carbon source. The use of
Na2CO3 as the base solution is therefore recommended
when the wastewater may have a deficiency of inorganic
carbon. This paragraph does not apply when titrimetry is
used.
• Synthetic medium
This should contain all the required macro- (sodium,
chloride, phosphate, magnesium, sulphate, calcium,
potassium) and micro-nutrients (iron, zinc, copper,
manganese, boron, molybdate, cobalt iodide) to ensure
that cells are not limited and in extreme cases to avoid the
failure of the test. Thus, although their concentrations
may seem very low, one must make sure that all of the
constituents are added to the solution in the required
amounts. Regarding the macro-nutrients, the following
composition (amounts per litre of the nutrient solution) is
recommended (based on Kampschreur et al., 2007): 72
mg NaH2PO4, 35 mg MgSO4·7H2O, 5 mg CaCl2·2H2O,
180 mg NaCl, 30 mg KCl, and 1 mg yeast extract. Micronutrients can be supplied by dosing 0.3 mL L-1 of a trace
element solution containing (per litre of solution, recipe
based on Kampschreur et al., 2007): 10 g EDTA, 1.5 g
87
FeCl3·6H2O, 0.15 g H3BO3, 0.03 g CuSO4·5H2O, 0.18 g
KI, 0.12 g MnCl2·4H2O, 0.06 g Na2MoO4·2H2O, 0.12 g
ZnSO4·7H2O, 0.15 g CoCl2·6H2O. Other similar nutrient
solutions can be used as long as they contain all the
previously reported required nutrients.
• Washing media
If the sludge sample must be washed to remove an
undesirable compound (which may be even inhibitory or
toxic), a washing media should be prepared. The same
synthetic medium described above can be used as a
washing medium. The washing process can be repeated
twice or three times applying the procedure described in
Section 2.2.3.5. Thereafter, the following preparation
steps of the batch activity tests can be performed. In
special cases when sludge from a full-scale plant is used,
the plant effluent may be used for washing purposes
(only if it does not contain toxic or inhibitory
compounds).
Prior to the execution of the experiment, samples of
the media and mixed liquor or activated sludge used to
perform the tests should be collected to confirm/check
the initial (desired) concentration of parameter(s) of
interest (e.g. ammonium, nitrite, nitrate, XTSS, XVSS).
Finally, the required working and stock solutions to
carry out the determination of the analytical parameters
of interest must also be prepared in accordance with
Standard Methods (APHA et al., 2012) and the
corresponding protocols.
2.4.5 Nitrification batch activity tests:
execution
Test NIT.CHE Nitrification chemical test: assessing the
maximum ammonium oxidation rate
Activated sludge preparation
1. For sample collection please refer to Section 2.4.3.4.
2. For conventional activated sludge samples from
wastewater treatment plants treating urban
wastewaters, a sample with a XVSS concentration of
around 2-4 g VSS L-1 would be suitable. XVSS
adjustment can be performed as suggested in Section
2.4.3.5.
3. Pour a defined volume (VML) of a mixed liquor
sample (typically 1 to 3 L) into the reaction vessel and
start mixing, aeration and the temperature and pH
control systems to maintain temperature and pH
around the desired set point values. Select a desired
set point pH value; typically values between 7.5 and
88
Execution of the test
1. Verify that the temperature, pH, and DO readings are
at the desired set point values or at least within the
selected intervals. Otherwise adjust the operating
conditions accordingly and wait until the system
stabilizes.
2. Once stable conditions are reached, add the
ammonium solution (having previously adjusted its
temperature to the target temperature of the test) to
achieve a neither limiting nor inhibiting ammonium
concentration in the mixed liquor. Typical and
adequate values are between 20 and 40 mg N L-1.
3. Start the stopwatch to keep precise track of the
sampling times, since formally the test starts with the
addition of the ammonium solution. Collect the
activated sludge samples every 20-30 min throughout
the execution of the test. Note that all samples need
to be filtered through 0.45 μm pore size filters (or
smaller), except those used for the determination of
XTSS and XVSS concentrations.
4. Conclude the test after 3 to 4 h or, if the sampling and
analytical determination of ammonium in the
collected samples allows it, when ammonium is
depleted.
5. After the conclusion of the test, take a sample for the
final XVSS assessment.
Note that nitrite is rather unstable; therefore, nitrite
concentrations have to be quickly assessed after sampling
in the same day (see Table 2.2.2).
Data analysis
A typical output of this test is outlined in Figure 2.4.3.
On the y-axis, the ammonium, nitrite and nitrate
concentrations (in mg N L-1) are reported, while time (in
hours) is reported on the X-axis. The linear regression
over these data allows for the assessment of the
ammonium removal and nitrate production rates (in mg
N L-1 h-1). Please note that the collected data should be
sufficient to reliably estimate the corresponding
removal/production rate (e.g. rNH4 and rNO3) by linear
regression with a satisfactory coefficient of
determination (e.g. R2 > 0.98). Thus, preferably at least 4
to 5 data points are needed to carry out the linear
regression, implying that a larger number of samples will
need to be collected in the beginning of the test.
35
NH4, NO2, NO3 (mgN L-1)
8.4 are adequate. If an automated pH control is not
available, correct the pH to the desired value by
manual addition of an acid/base solution. In principle,
the closer the pH set point value to the typical
operational pH at the plant, the closer and more
representative the nitrification rate will be. The same
principle applies to the selection of the temperature
set point value. As for DO, the aeration system
should provide sufficient oxygen to avoid DOlimiting conditions during the execution of the
nitrification tests. This means that under endogenous
conditions the observed DO value should be high
(e.g. > 6 mg L-1).
4. Wait for approximately 30 min to reach and ensure
stable initial conditions. This pre-incubation phase
will normally allow any residual nitrite remaining
from the plant or source of origin of the sludge to be
consumed.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
30
25
rNH4
20
rNO3
15
10
5
0
0
0.5
1.0
1.5
Time (h)
2.0
2.5
3.0
Figure 2.4.3 Typical profiles obtained in a Test NIT.CHE: ammonium (●),
nitrite (● ) and nitrate (● ) concentrations are displayed on the y-axis.
Relevant rates of interest are also displayed (e.g. ammonium removal rate
rNH4, and nitrate production rate rNO3).
Nitrite may accumulate up to few mg N L-1 when
conventional activated sludge is used. However, if the
activated sludge is used to perform (partial) nitritation
tests, then a higher nitrite accumulation will be expected
and its concentration should be monitored in time, similar
to ammonium and nitrate (for example see Test
NIT.TIT.2). When using conventional activated sludge to
perform the full oxidation of ammonium to nitrate, the
ammonium removal rate should equal the nitrate
production rate with a negligible accumulation of nitrite
during the test. The maximum specific ammonium
oxidation rate (qAOO,NH4, as mg N g VSS-1 h-1) can be
therefore computed as follows:
qAOO,NH4 = rNH4 / XVSS
Eq. 2.4.24
ACTIVATED SLUDGE ACTIVITY TESTS
Test NIT.TIT.1 Nitrification titration test: assessing the
maximum ammonium oxidation rate
Activated sludge preparation
1. For conventional activated sludge samples from
wastewater treatment plants treating urban
wastewaters, a sample with a XVSS of around 2-4 g
VSS L-1 will be suitable. For very diluted or
concentrated sludge samples, a concentration step
(e.g. by decanting into an Imhoff cone for 30 min or
by centrifuging at 4,000 rpm for a few minutes) or
dilution with secondary effluent from the same
wastewater treatment plant may be helpful. This will
avoid having too slow or too fast nitrification rates.
2. Pour a known volume of the activated sludge sample
(typically 1 L) into the reaction vessel and activate
the aeration and the temperature control systems.
Select a desired set point pH value. Typically, values
between 7.5 and 8.4 are adequate. In principle, the
closer the set point pH value to the typical operational
pH of the plant or source of sludge, the closer and
more representative the observed nitrification rate
will be. The same principle applies to the selection of
the temperature value. Concerning DO, the aeration
system should provide sufficient oxygen to avoid
DO-limiting conditions during the course of the
nitrification tests. This means that under endogenous
conditions the observed DO value should be
relatively high (e.g. > 6 mg L-1).
3. Activate the automated titration system for a preincubation period of approximately 1 h. This preincubation phase will ensure that: (i) endogenous
conditions are achieved at the start of the titration test
(SB, ammonium and nitrite are oxidized during this
overnight aeration phase) and (ii) that temperature,
pH and SO2 are stable at the start of the test. Prehumidified air may be used to limit significant water
evaporation during this pre-incubation phase. Note
that prolonged incubation periods (e.g. longer than 4
h) may reduce the nitrification rate due to fast
endogenous biomass decay under aerobic conditions.
Execution of the test
1. Activate the data logging.
2. Add the ammonium chloride stock solution (having
previously adjusted its temperature to the target
temperature of the test) to achieve an ammonium
concentration in the activated sludge that is neither
limiting nor inhibiting (between 20 and 40 mg N L-1
are typically adequate values). NH4Cl addition to an
alkaline suspension has an acidifying affect (acid
hydrolysis) that leads to a rapid pH drop that should
89
be compensated by an automatic pH control system
or through the manual addition of concentrated
NaOH. Any addition of the titration solutions during
this pH-adjustment phase should be disregarded
during the data analysis. Upon the ammonium
addition, nitrifying bacteria will oxidize the
ammonium added and consequently an alkaline
titrating solution will need to be added to compensate
for the acidifying nitrification effect.
3. Record the volume of the NaOH titration solution
added over time (VNaOH versus time) (20-40 min are
usually adequate). Check that the pH reading value
remains close to the target pH set point ± 0.02 and
that the SO2 level does not become limiting. Do not
change the aeration rate or the mixing conditions
since this would affect the titration rate, making the
assessment of the nitrification-related titration rate
cumbersome. The collected data should be sufficient
to reliably estimate the titration rate (QNaOH) from a
linear regression of VNaOH versus time data with a
satisfactory coefficient of determination (R2 > 0.98).
4. Add allyl-N-thiourea in order to achieve a final ATU
concentration of 10 mg L-1. At this concentration,
ammonium oxidation will be inhibited. Continue
recording the NaOH titration rate for a further 20-30
min in order to assess the residual titration rate due to
background pH affection reactions such as CO2
stripping (QNaOH,final), if present. Note that the longer
the pre-incubation period the lower the relevance of
QNaOH,final.
5. The test can now be ended. Measure the final
activated sludge volume and take a sample to assess
the XVSS concentration. Note that the suspension
volume will change during the course of the test due
to the addition of titration solutions. It is expected that
the addition of titration solution does not account for
more than 10 % of the final activated sludge volume.
Data analysis
A typical trend of the volume of NaOH added during the
test is depicted in Figure 2.4.4. From these data (titration
volume versus time data), titration rates (Q) can be
computed through the slope of the titration curve using
the corresponding tool in a worksheet or by applying the
following formula:
Q=
n · ∑ ti · VNaOH,i ‒ ∑ ti · ∑ VNaOH,i
n · ∑ t2i ‒ ( ∑ ti )
2
Eq. 2.4.25
Where, n is the number of recorded data [ti, VNaOH,i]
available.
90
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2.
ATU
VNaOH
QNaOH,final
NH4Cl
QNaOH
Time
Figure 2.4.4 Example of a pH-static titration curve during a nitrification
test to assess the maximum ammonium-oxidizing capacity. Arrows
indicate the addition of ammonium chloride and allyl-N-thiourea. Relevant
titration rates are also identified on the graph.
3.
The NaOH titration rates (QNaOH and QNaOH,final in mL
min-1) are used to assess the ammonium oxidation rate
(FNHx in mg N min-1) by taking into account the
concentration of the NaOH solution (NNaOH in meq mL-1)
and the ratio between ammonium oxidation and
alkalinity consumption YNH4_H+ that can be assessed from
Eq. 2.4.15. Therefore:
4.
FNHx = (QNaOH ‒ QNaOH,final ) · NNaOH ·YNH4_H+
Eq. 2.4.26
Finally, the maximum specific ammonium oxidation
rate of the sludge (qAOO,NH4, in mg N g VSS-1 h-1) can be
computed by taking into account the XVSS concentration
of the sludge sample (in g VSS L-1) and the suspension
volume observed at the end of the test (VML):
qAOO,NH4 = 60 · FNHx/ (VML · XVSS)
5.
Eq. 2.4.27
Test NIT.TIT.2 Nitrification titration test: assessing the
maximum ammonium and nitrite oxidation rates
Activated sludge preparation
Follow steps 1, 2 and 3 of the activated sludge
preparation described for Test NIT.TIT.1.
Execution of the test
1. Select an appropriate set point value for SO2 (DO set
point). Typically values between 4.0 mg N L-1 and 6.0
6.
mg N L-1 are adequate to assess the maximum
nitrification rates. Activate the data-logging system.
Record the volumes of the H2O2 and NaOH titration
solutions added over time (VH2O2 and VNaOH versus
time) (20-40 min are normally adequate). Check that
the pH and DO values remain within the interval pH
set point ± 0.02 and SO2 set point ± 0.10 mg L-1,
respectively. The collected data should be sufficient
to reliably estimate (i.e. with a satisfactory coefficient
of determination, R2 > 0.98) the alkaline titration rate
from the linear regression of VNaOH versus time data,
and the oxygen titration rate (QH2O2) from the linear
regression of VH2O2 versus time data. During this
phase, titration rates are triggered by endogenous
respiration which leads to DO consumption and CO2
production; the former is compensated by H2O2
addition (at a rate indicated as QH2O2,ini), and the latter
by NaOH addition (at a rate indicated as QNaOH,ini).
Add nitrite at a neither limiting nor inhibiting nitrite
concentration in the activated sludge (around 10 mg
N L-1 are typically adequate values) in order to trigger
nitrite oxidation. Repeat data acquisition as described
in Step 2 in order to estimate the oxygen titration rate
that includes the oxygen request for nitrite oxidation
(QH2O2,NO2). The alkaline titration rate will not change
since nitrite oxidation does not significantly affect the
suspension pH.
Add the ammonium chloride stock solution according
to the instructions reported in step 2 of the test
operation procedure described for Test NIT.TIT.1.
This addition will trigger ammonium oxidation as
well. Repeat data acquisition as described in step 2 to
assess the alkaline titration rate (QNaOH,NH4) and the
oxygen titration rate (QH2O2,NH4) that include the
ammonium oxidation needs.
Add allyl-N-thiourea (ATU) to a final concentration
of 10 mg L-1. Ammonium oxidation will be inhibited.
Continue recording the NaOH titration rate for a
further 20-30 min to assess the residual titration rate
due to the background pH affecting reactions such as
CO2 production (QNaOH,final) and oxygen-affecting
including
endogenous
reactions
(QH2O2,final)
respiration and residual nitrite oxidation.
End the test according to the instructions reported in
Step 4 of the test operation procedure described for
Test NIT.TIT.1.
Data analysis
A typical trend of the cumulated volume of titration
solutions added during the test is depicted in Figure 2.4.5.
From these data (volume versus time data), the titration
ACTIVATED SLUDGE ACTIVITY TESTS
91
rates (Q) can be computed as the slope of the titration
curve using the corresponding tool in a worksheet or by
applying the formula previously described in Test
NIT.TIT.1.
VNaOH, VH2O2
ATU
QH2O2,final
NH4Cl
QNaOH,final
QH2O2,NH4
NaNO2
QNaOH,NH4
QH2O2,NO2
2
tNH4
YNH4/O2_NO3 =
1
= 0.23 g N g O-1
2
4.57 ‒ YAOO
tATU
The NaOH titration rates (QNaOH,NH4 and QNaOH,final in
mL min-1) can be first used to assess the ammonium
oxidation rate (FNHx,NaOH in mg N min-1) by taking into
account the concentration of the NaOH titration solutions
(NNaOH in meq mL-1) and the ratio between ammonium
oxidation and alkalinity consumption, as suggested in
Test NIT.TIT.1:
FNHx,NaOH = (QNaOH, NH4 ‒QNaOH, final ) · NNaOH ·YNH4_H+
Eq. 2.4.28
Similarly, oxygen titration rates (QH2O2,NH4 and
QH2O2,final in mL min-1) can be first used to assess the
ammonium oxidation rate (FNHx,H2O2 in mg N min-1) by
taking into account the concentration of the H2O2 titration
solutions (NH2O2 in mmol O2 mL-1) and the ratio between
ammonium oxidation to nitrate and oxygen consumption,
YNH4/O2_NO3, that is, according to the ASM nitrification
stoichiometry (Henze et al., 2000):
Eq. 2.4.30
The oxygen titration rates collected during step 3
(QH2O2,NO2 in mL min-1) will be used to determine the
nitrite oxidation rate (FNO2 in mg N min-1) by taking into
account the ratio between nitrite oxidation to nitrate and
oxygen consumption, YNO2/O2_NO3, as follows (according
to the two-step nitrification stoichiometry):
Time
Figure 2.4.5 Example of a pH/DO-stat titration curve during a nitrification
test executed to assess the maximum ammonium- and nitrite-oxidizing
capacity. Arrows indicate the addition of nitrite, allyl-N-thiourea and
ammonium. Relevant titration rates of interest are also displayed on the
graph.
Eq. 2.4.29
The values of FNHx,NaOH and FNHx,H2O2 should be
similar and their comparison can be used to validate the
experimental data. Differences higher than 15 % may
suggest the need for a careful verification of the
experimental setup.
FNO2 = (QH2O2,NO2 ‒ QH2O2,initial )
· NH2O2 · 32 · YNO2/O2_NO3
QH2O2,initial ,QNaOH,initial
tNO
FNHx, H2O2 = (QH2O2,NH4 ‒ QH2O2,final )
· NH2O2 · 32 · YNH4/O2_NO3
YNO2/O2_NO3 =
1
= 0,88 g N g O-1
2
1.14
Eq. 2.4.31
Eq. 2.4.32
If no difference is observed between QH2O2,NO2 and
QH2O2,initial values then either the nitrite oxidation rate is
very slow or it is much slower than the ammonium
oxidation rate. This may lead to nitrite accumulation
during the endogenous phase since ammonium oxidation
would probably take place on the ammonium released
through ammonification. If so, QH2O2,initial would include
oxygen consumption due to the slow nitrite oxidation
process and be equal to QH2O2,NO2. In this case, the
activated sludge sample should be elutriated to remove
any nitrite content by centrifugation and resuspension in
a nitrite-free physiological medium to assess the nitrite
oxidation rate. However, when the nitrite oxidation rate
is very slow, chemical tracking over a longer testing
period (a few hours) may lead to more reliable estimates
and should be preferred. Note that, for very slow nitrite
oxidation rates, FNHx,H2O2 may be higher than FNHx,NaOH.
As a matter of fact, the use of YNH4/O2_NO3 is no more
correct since the sole ammonium oxidation to nitrite
request should be taken into account, since YNH4/O2_NO3
quantifies the overall ammonium plus nitrite oxidation
request. Therefore, in such a case, FNHx,H2O2 would be
more correctly estimated taking into account the ratio
between ammonium oxidized and oxygen consumption
for ammonium oxidation to nitrite, YNH4/O2_NO2, which
92
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
can be expressed according to the two-step nitrification
stoichiometry expression as:
carbon solutions that require specialized metabolic
capabilities/microorganisms.
FNHx,H2O2 = (QH2O2,NH4 ‒ QH2O2,final )
· NH2O2 · 32 · YNO2/O2_NO2
Four tests are presented. The first one (DEN.CHE.1)
refers to the use of an easily biodegradable carbon source,
for which both the denitrification rate and the biomass
anoxic growth yield are relevant parameters to be
assessed. In the second test (DEN.CHE.2), a real
wastewater is used. Although kinetic information can be
drawn from this, this second test is mainly meant to
assess the denitrification capacity of this wastewater, i.e.
the amount of nitrate that can be denitrified per unit
volume of this specific wastewater. Finally, two more
tests are presented for the assessment of the maximum
denitrification rate by applying a manometric
(DEN.MAN) or titrimetric (DEN.TIT) tracking
procedure.
YNO2/O2_NO2 =
Eq. 2.4.33
1
= 0.31 g N g O-1
2
3.43 ‒ YAOO
Eq. 2.4.34
Moreover, the difference between QNaOH,initial and
QNaOH,final makes it possible to estimate the
ammonification rate under endogenous conditions. The
first titration rate compensates for the alkaline effect of
endogenous respiration and of nitrification, which is
limited by the ammonification process responsible for
ammonium release. The ammonification-related
alkalizing effect is no longer present after ATU addition.
Thus, the ammonification rate (FN_NHx in mg N min-1) can
be estimated, by difference, according to the following
equation:
FN_NHx = YNH4/O2_NO3 · (QNaOH,initial ‒ QNaOH,final ) · NNaOH
Eq. 2.4.35
Finally, the maximum biomass-specific ammonium
and nitrite oxidation rates of the sludge (qAOO,NHx, and
qNOO,NO2_NO3 in mg N g VSS-1 h-1) and the specific
ammonification rate (qN_NHx), can be computed by taking
into account the XVSS concentration of the sludge sample
(in g VSS L-1) and the suspension volume observed at the
end of the test (VML):
qAOO,NHx = 60 · FNHx / (VML · X
)
qNOO,NO3_NO2 = 60 ·FNO2 / (VML · X
qN_NHx = 60 · FN_NHx / (VML · X
)
Eq. 2.4.27
)
Eq. 2.4.36
Eq. 2.4.37
2.4.6 Denitrification batch activity tests:
preparation
These tests are meant to assess the maximum
denitrification rate of a sludge sample and the anoxic
biomass growth yield. Various types of carbon sources
can also be used such as internal carbon, external carbon
sources (e.g. sugar or alcohol) or wastewater. Typically,
the maximum denitrification rate is expressed when a
rapidly biodegradable carbon source (to which the sludge
is adapted) is used, while lower rates are observed in the
presence of complex organic molecules that require a
preliminary hydrolysis step or when dosing external
2.4.6.1 Apparatus
Each tracking methodology (chemical, titrimetric or
manometric) has special apparatus requirements.
When the chemical tracking is applied refer to the
following list of equipment:
1. An (airtight) batch reactor equipped with mixing
system and adequate sampling ports (as described in
Section 2.4.3.1).
2. A nitrogen gas supply (recommended).
3. A calibrated pH electrode (if not included or
incorporated in the batch reactor setup).
4. A 2-way pH controller for HCl and NaOH addition
(alternatively a one-way control, generally for HCl
addition, or a manual pH control can be applied
through the manual addition of HCl and NaOH).
5. A thermometer (with a recommended working
temperature range of 0 to 40 °C).
6. A temperature control system (if not included in the
batch reactor setup).
7. A DO meter with an electrode (if not
included/incorporated in the batch reactor setup) to
verify anoxic conditions.
8. A stopwatch.
When titrimetric tests are performed, refer to the
following list of equipment:
1. The equipment for the titrimetric system described in
Section 2.4.3.1.
2. A nitrogen gas supply (recommended).
When manometric tests are performed, refer to the
following list of equipment:
ACTIVATED SLUDGE ACTIVITY TESTS
1. The equipment for the manometric system described
in Section 2.4.3.1.
2. A nitrogen gas supply (recommended).
2.4.6.2 Materials
For a complete list of required materials refer to Section
2.2.3.2 (with the exception of points 7 and 8).
2.4.6.3 Working solutions
• Real wastewater
For batch tests that require the use of real wastewater,
follow the instructions given in Section 2.2.3.3.
93
fact that their concentrations may seem very low, one
must make sure that all of the constituents are added to
the solution in the required amounts. Regarding macronutrients the following composition (amounts per litre of
nutrient solution) is recommended (based on Smolders et
al., 1994): 107 mg NH4Cl, 90 mg MgSO4·7H2O, 14 mg
CaCl2·2H2O, 36 mg KCl, 1 mg yeast extract. Micronutrients can be supplied by dosing 10 mL L-1 of a trace
element solution containing (per litre of solution) (based
on Vishniac and Santer, 1957): 50 g EDTA, 22 g
ZnSO4·7H2O, 5.54 g CaCl2, 5.06 g MnCl2·4H2O, 4.99 g
FeSO4·7H2O, 1.10 g (NH4)6Mo7O24·4H2O, 1.57 g
CuSO4·5H2O, and, 1.61 g CoCl2·6H2O. Similar nutrient
solutions can be used as long as they contain all the
previously reported required nutrients.
• Carbon source solution
This is usually composed of a readily biodegradable
carbon source (SB), preferably volatile fatty acids like
acetate or propionate, sugars, or alcohol solutions. The
choice of the organic carbon source depends on the
nature or goal of the test and the corresponding research
questions. Sometimes more complex substrates are used
that are more similar to real wastewaters, containing a
mixture of readily and slowly biodegradable COD. For
anoxic batch activity tests, the COD concentration (both
total and soluble) should be known in order to select a
proper dose.
It is recommended to take a sample of the media and
sludge prior to the execution of the experiment to
confirm/check the initial (desired) concentration of
parameter(s) of interest (e.g. COD, nitrate/nitrate, XVSS).
• Nitrate or nitrite solutions
2.4.6.4 Material preparation
Nitrate and nitrite salts are needed to adjust the
nitrate/nitrite level during denitrification tests.
• Washing media
If the sludge sample must be 'washed' to remove
undesirable compounds, then refer to Section 2.4.3.4.
• Acid and base solutions
These are 100-250 mL of 0.2 M HCl and 100-250 mL 0.2
M NaOH solutions for automatic or manual pH control,
and 10-50 mL of 1 M HCl and 10-50 mL 1 M NaOH
solutions for an initial pH adjustment if the desired
operational pH is very different from the pKa value of the
buffering agent. For titrimetric tests, titration solutions
should be prepared according to the instructions given in
Section 2.4.4.3.
• Nutrient solution
This should contain all the required macro- (ammonium,
magnesium, sulphate, calcium, potassium) and micronutrients (iron, zinc, calcium, copper, manganese,
molybdate, cobalt) to ensure that cells are not short of
basic nutrients for their metabolism. Thus, despite the
Finally, the required working and stock solutions to
carry out the determination of the analytical parameters
of interest must be also prepared in accordance to
Standard Methods (APHA et al., 2012) and the
corresponding protocols for their preservation and
analytical determination.
For general instructions on how to organize the material
preparation, please refer to Section 2.2.3.4 and to Table
2.2.2. Test-specific requirements will be listed within
each test protocol.
2.4.7 Denitrification batch activity tests:
execution
Test DEN.CHE.1 Denitrification chemical test: assessing the
maximum denitrification rate and the anoxic growth yield in
the presence of a specific carbon source
Activated sludge preparation
The optimal sampling point for the activated sludge
would be the outlet of the post-denitrification tank.
1. For conventional activated sludge samples treating
urban wastewaters, a sample with a XVSS of around 24 g VSS L-1 would be suitable. The XVSS can be
adjusted as suggested in Section 2.4.3.5.
2. Pour a known volume (VML) of activated sludge
sample (typically 1 to 3 L) into the reaction vessel and
94
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
4.
5.
6.
Execution of the test
1. Add the nitrate and SB stock solutions. The
temperature of both these solutions should have been
previously adjusted to the target temperature of
interest. The initial nitrate concentration should be
neither limiting nor inhibitory (20 to 25 mg N L-1 are
typically adequate values). If residual nitrate is
expected to be present in the activated sludge sample,
the nitrate addition should then be reduced.
2. Add a non-limiting amount of the readily
biodegradable carbon source (SB). To assess the
appropriate amount of SB, one can consider the
stoichiometric relationship between the amount of SB
and nitrate consumed during denitrification:
YNO3_SB,Ax =
2.86
(g COD g N-1 )
1 ‒ YOHO,Ax
7.
It should be noted that nitrite concentration is
unstable, and therefore nitrate and nitrite concentrations
have to be quickly measured, preferably on the same day.
Data analysis
The typical output of this test is outlined in Figure 2.4.6.
rNOx_N2,endo
r COD
SNO3,eq
Verify that your target values of temperature, pH, and SO2
are close to the set points or within the selected intervals.
Otherwise adjust and wait until the system stabilizes.
may lead to biomass inhibition. An ammonium stock
solution can also be added to adjust the ammonium to
SB ratio to 0.05 g N g COD-1.
The test starts with the addition of the nitrate and SB
solutions. Start the stopwatch to keep precise track of
the following sampling times and start the sampling
campaign.
Collect activated sludge samples at regular time
intervals. As a general suggestion, samples for the
determination of the C source (or of soluble COD
depending upon the analytical parameter of interest)
and of nitrite and nitrate must be collected every 10
min in the first 30 min after SB addition, every 15 min
during the following 60 min, and later on every 30
min until the end of the test.
Conclude the test when the nitrate and nitrite are fully
depleted. Use nitrate and nitrite strip tests to quickly
assess when these compounds have been consumed.
Take a sample for final XVSS concentration
assessment.
rNOx_N2,exo
Eq. 2.4.38
3. The SB addition should guarantee that the SB to nitrate
ratio should be at least twice as much as the
stoichiometric value found with the previous
expression. Note that the anoxic biomass growth
yield, YOHO,Ax, depends on the carbon source, as
reported in Table 2.4.2. However, a value of 0.5 can
be typically used for a rough estimation. Under this
assumption, a SB concentration in the activated
sludge of 200 mg COD L-1 would usually be
adequate. This concentration also satisfies the initial
SB to XVSS ratio value (0.05-0.1 g COD g VSS-1)
suggested in Section 2.2.4.1. Lower values may result
in too rapid carbon depletion while too high values
COD
start mixing. Also start the temperature and pH
control systems to keep both parameters around the
desired set point values.
3. Sparge N2 into the reaction headspace for
approximately 10 min to ensure a deoxygenated
environment. Ensure a gas outlet to limit
overpressure. Gas sparging can be continued until the
end of the test. If this is not feasible, then a proper
airtight reactor with a gas outlet preventing oxygen
back-diffusion (e.g. by using a unidirectional check
valve or a water lock) should be used (see Section
2.2.2.1 for more information on how to ensure anoxic
conditions).
4. Wait for approximately 30 min to ensure stable initial
conditions. This pre-incubation phase will usually
allow the removal of any residual nitrate.
NO3
RBCOD
Time
Figure 2.4.6 Typical profiles obtained in a Test DEN.CHE.1: nitrate
concentrations (●) on the main y-axis, COD concentrations (●) on the
secondary y-axis. Relevant rates of interest are also displayed (endogenous
denitrification rate rNOx_N2,endo, exogenous denitrification rate, rNOx_N2,exo, and
COD consumption rate rCOD). The arrow indicates the substrate addition.
ACTIVATED SLUDGE ACTIVITY TESTS
95
On the principal y-axis, the oxidized nitrogen
equivalent SNO3,Eq is reported, which corresponds to a
weighted sum of the nitrate and nitrite concentrations:
Eq. 2.4.39
SNO3,Eq = SNO3 + 0.6 · SNO2
The 0.6 weight applied to the nitrite concentration
corresponds to the relative electron-accepting capacity of
nitrite with respect to nitrate (1.71/ 2.86 = 0.6), as
suggested by Kujawa and Klapwijk (1999). On the
secondary y-axis, soluble COD data are presented.
During the first period (i.e. before the addition of the
external carbon source), endogenous denitrification takes
place and a slow-rate reduction of SNO3,Eq is observed.
The linear regression over these data (see Section
2.4.4.2) can be used to assess the endogenous
denitrification rate (rNOx_N2,endo, in mg N L-1 min-1). After
the SB addition, the availability of the exogenous carbon
source speeds up the consumption of nitrate and nitrite.
Collecting the nitrate/nitrite data afterwards, but before
nitrate/nitrite becomes limiting, makes it possible to
assess the exogenous denitrification rate (rNOx_N2,exo, in
mg N L-1 min-1). Similarly, the maximum COD
consumption rate (rCOD, in mg COD L-1 min-1) can be
assessed within the same timeframe.
The maximum specific denitrification rate on SB
(qNOx_N2,SB in mg N g VSS-1 h-1) on the tested carbon
source can be computed as follows:
qNOx_N2,SB = 60 · (rNOx_N2,exo ‒ rNOx_N2,endo ) /XVSS
Eq. 2.4.40
Moreover, by combining denitrification rates and the
COD consumption rate, the biomass growth yield
(YOHO,Ax) can also be assessed according to the following
formula:
YOHO,Ax = 1 ‒ 2.86
(rNOx_N2,exo ‒ rNOx_N2,endo )
rCOD
Eq. 2.4.41
Test DEN.CHE.2 Denitrification chemical test: assessing the
denitrification potential of wastewater
Activated sludge preparation
Samples for this test should be collected at the outlet of
the pre-denitrification tank.
1. Follow step 1 of the activated sludge preparation
procedure described for Test DEN.CHE.1 Note that
in this case a biomass concentration of 3-4 g VSS L-1
would be more adequate since a dilution effect is
obtained when the wastewater is added.
2. Pour a known volume of the activated sludge sample
(VML, typically 0.6-0.8 L) and keep a sample of it to
assess the volatile suspended solids concentration
(MLVSS in g VSS L-1). Follow steps 2, 3 and 4 of the
activated sludge preparation procedure described for
Test DEN.CHE.1.
Test execution
1. Verify that your target values of temperature, pH, and
DO are within the selected intervals. Otherwise adjust
them and wait for their stabilization.
2. Select the appropriate volume of wastewater (VWW)
to be added. It would be ideal to add an amount of
wastewater so that the final biodegradable COD
concentration in the reaction vessel remains within 30
and 70 mg L-1. By assuming a typical concentration
of biodegradable carbon (SB + XCB) of 100-180 mg
L-1, a dilution factor (VML:VWW) of 2 to 6 should be
appropriate. Pour the wastewater into the reaction
vessel and add the nitrate stock solution in order to
achieve an initial nitrate concentration in the final
mixture (VML+VWW) of 20-25 mg N L-1.
3. Start a stopwatch in order to keep precise track of the
following sampling times and start the sampling
campaign. As a general suggestion, samples for the
determination of nitrite and nitrate concentration
must be collected every 5 min during the first 30-45
min of execution of the test, every 10 or 15 for a
further 30-45 min, and later on every 15 or 30 min
until the end of the test.
4. Conclude the test after 3-4 h when a slow
(endogenous-like) rate is observed.
Data analysis
The typical output of this test is outlined in Figure 2.4.7.
On the y-axis, the oxidized nitrogen equivalent SNO3,Eq
during the course of the test is given. Various
denitrification rates can be observed. The highest rate
(rNOx_N2,SB) is observed initially, i.e. when both SB and
XCB are available to the heterotrophic microorganisms
(interval Δt1 in the graph). Once the SB organic fraction
is fully utilized, denitrification proceeds only on the
soluble organics that are made available through the
hydrolysis of XCB, therefore on the so-called slowly
biodegradable organic material (interval Δt2 in the
graph). The denitrification rate (rNOx_N2,XCB) is therefore
limited by the hydrolysis rate of XCB. When hydrolysable
organics are fully consumed, denitrification can continue
on the endogenous carbon at a slower rate (rNOx_N2,endo) as
long as nitrate and nitrite are still available.
96
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
∆t1
∆t2
r NOx_N2,SB
SNO3,eq
SNO3/SB,eq
SNO3/XCB,eq
By considering the volume of wastewater tested, the
denitrification potential of the rapidly biodegradable
(DPSB) and slowly biodegradable (DPXCB) organic
compounds can be finally assessed:
DPSB =
SNO3/SB,eq · (VML + VWW )
VWW
rNOx_N2,XCB
DPXCB =
Nitrate
Wastewater
Activated sludge preparation
Figure 2.4.7 Typical profiles obtained in a DEN.CHE.2 test. Nitrate
concentrations during the execution of the test are displayed ( ● ).
Relevant rates of interest are shown (denitrification rate on the rapidly
biodegradable rNOx_N2,SB, and slowly biodegradable rNOx_N2,XCB, fractions and
endogenous denitrification rate rNOx_N2,endo). Arrows indicate the nitrate and
wastewater additions. Relevant nitrate equivalent variations are also
identified (SNO3/SB,eq on SB and SNO3/XCB,eq on XCN).
The linear regression of SNO3,Eq versus time data for
each interval (see Section 2.4.4.2) allows each relevant
denitrification rate (in mg N L-1 min-1) to be calculated.
Specific denitrification rates (in mg N g VSS-1 h-1) can be
computed as follows:
specific denitrification rate on SB:
Eq. 2.4.42
specific denitrification rate on XCB:
qNOx_N2,XCB = 60 · (rNOx_N2,XCB ‒ rNOx_N2,endo ) / XVSS
Eq. 2.4.43
•
Patm
MN2
(273 + T )
·
· 22.4 ·
VHS 28 · 1,000
273
Eq. 2.4.49
Where, TC is the temperature (°C), MN2 is the
mass of nitrogen gas generated by denitrification
during the course of the test (mg N), which depends
on the nitrate concentration (SNO3_N2,Ax, in mg N L-1)
and the activated sludge volume, as follows:
MN2 = SNO3_N2,Ax · VML
specific endogenous denitrification rate:
qNOx_N2,endo = 60 · (rNOx_N2,endo ) / XVSS
1. Follow step 1 of the activated sludge preparation
protocol described for Test DEN.CHE.1.
2. Select the appropriate amount of activated sludge to
be poured into the reaction vessel (VML). For this
purpose, one should consider that the extent of the
overpressure caused by gas release depends on the
remaining headspace volume (VHS) and on the
expected N2 generation, (amount of nitrogen to be
denitrified) and the headspace volume (VHS). Hence,
the correct selection of this ratio is crucial to avoid
extreme pressures (either too high or too low). The
maximum overpressure will be achieved at the end of
the test, i.e. when all the nitrate has been denitrified,
described as:
Pmax ‒ Patm =
qNOx_N2,SB = 60 · (rNOx_N2,SB ‒ rNOx_N2,endo ) / XVSS
•
Eq. 2.4.48
Test DEN.MAN Denitrification manometric test: assessing the
denitrification kinetic rate
rNOx_N2,endo
Time
•
SNO3/XCB,eq · (VML + VWW )
VWW
Eq. 2.4.47
Eq. 2.4.44
The amount of nitrate equivalents that are consumed
on the rapidly (SNO3/SB,eq) and the slowly (SNO3/XCB,eq)
biodegradable carbon sources are also illustrated in the
graph which can be estimated as follows:
SNO3/SB,eq = (rNOx_N2,SB ‒ rNOx_N2,XCB ) · ∆t1
Eq. 2.4.45
SNO3/XCB,eq = (rNOx_N2,XCB ‒ rNOx_N2,endo ) · ∆t2
Eq. 2.4.46
Eq. 2.4.50
By substituting this equation into the previous
one and rearranging, the following relationship is
obtained:
VML Pmax ‒ Patm
28
273
1
=
·
·
·
VHS
Patm
22.4 (273 + TC ) SNO3_N2,Ax
1,000
Eq, 2.4.51
The optimal value for Pmax is normally around 0.2
atm, which means that, under typical conditions (TC
ACTIVATED SLUDGE ACTIVITY TESTS
~20 °C, SNO3_N2,Ax ~20 mg N L-1), a VML to VHS ratio
of around 11 is obtained. Therefore, for a total reactor
volume of 1 L, the ideal activated sludge volume to
be used during the test will be 0.92 L.
3. Pour the previously computed amount of activated
sludge into the reaction vessel. Insert the magnet for
the mixing. Insert NaOH pellets in the headspace for
CO2 adsorption. Flush the reactor headspace with N2
and seal the bottle gas tightly. Place the reactor into
the thermostatic chamber and under gentle mixing,
wait for 30 min until the temperature stabilizes.
Test execution
1. Determine the volume of nitrate solution to be added
(as a pulse or spike) to achieve a nitrate concentration
in the activated sludge of 20-25 mg N L-1. Using a
syringe, inject the nitrate stock solution through the
rubber septum. Select the amount of carbon source to
be dosed (also as a pulse or spike) in order to operate
under non-limiting and non-inhibiting conditions (see
step 4 of the execution of Test DEN.CHE). Then
inject the carbon source solution and follow up the
time execution with a stopwatch.
2. Start the manometric data collection. Collect data
every 15-30 min or until a bending point is observed
in the overpressure curve, which indicates the
exhaustion of nitrate.
3. End the test. Check the final pH and take a final
sample to measure the MLVSS concentration.
Data analysis
P (t)
A typical output of a manometric denitrification test
using a manometer with a data logger is outlined in
Figure 2.4.8.
rP
97
During the first 10-15 min, various phenomena may
overlap caused by potential interferences such as: (i)
intrusion into the bulk liquid of residual oxygen that may
remain in the headspace, (ii) water vapour pressure
equilibrium. (iii) initial N2 accumulation in the liquid
phase, and/or, (iv) microbiological lag phases. For this
reason, the initial overpressure data should be
disregarded. The following overpressure data can be used
to compute the pressure production rate (rP, atm min-1) by
linear regression (see Section 2.4.3.3).
From rP, the denitrification rate (FNO3_N2, mg N min-1)
can be assessed as follows:
FNO3_N2 =
rp
28
273
· VHS ·
·
Patm
22.4 273 + T
Eq. 2.4.52
With, Patm in atm and VHS in mL.
Finally, the specific denitrification rate qNOx_N2 (mg N
g VSS-1 h-1) can be computed using the activated sludge
MLVSS concentration (XVSS):
qNOx_N2 = 60 · FNO3_N2 / (VML · XVSS )
Eq. 2.4.53
Test DEN.TIT Denitrification titrimetric test: assessing the
denitrification kinetic rate
Activated sludge preparation
For conventional activated sludge samples from
wastewater treatment plants treating urban wastewaters,
a sample with XVSS of around 2-4 g VSS L-1 will be
suitable. The sample should be collected at the outlet or
end of the pre-denitrification or post-denitrification tank,
depending on the target treatment section.
1. Pour a known volume of the activated sludge sample
(typically 1 L) into the reaction vessel and start the
temperature control systems. Select the desired pH
and temperature set point values.
2. Start the automatic titration system and leave the
activated sludge under these conditions for a preincubation period (ideally 1 h). The pre-incubation
phase will encourage the consumption of any residual
nitrate or nitrite remaining from the plant.
Test execution
Time
Figure 2.4.8 Typical overpressure profile obtained in a manometric
denitrification test performed with a manometer equipped with a data
logger. The chart displays the maximum pressure production rate rP.
1. Start the data logging.
2. Add the nitrate stock solution (after adjusting its
temperature to the target temperature of the test) to
achieve a non-limiting nitrate concentration in the
activated sludge (between 10 and 20 mg N L-1 are
typically adequate values) and the SB solution at a
98
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
concentration that is neither limiting nor inhibiting.
Note that the amount of nitrate added (MNOx,ini in mg
N) has to be known. To assess the appropriate amount
of SB, the stoichiometric relationship between the
amount of SB and nitrate consumed can be used,
which can be calculated as follows:
2.86
(g COD g N-1 )
1 ‒ YOHO,Ax
Eq. 2.4.38
Note that the same type of test can be performed
using nitrite instead of nitrate. In this case, the
following relationship should be applied:
1.71
YNO2_SB,Ax =
(g COD g N-1 )
1 ‒ YOHO,Ax
Eq. 2.4.54
The SB addition can therefore be calibrated in
order to guarantee that the SB to nitrate (or nitrite)
ratio is at least 3-4 times the stoichiometric value
reported above. Note that YHD depends on the carbon
source. Nevertheless, a value of 0.5 can be typically
used for this rough estimation. This means that the SB
addition should be calibrated in order to ensure a SB
concentration in the activated sludge of 350 mg COD
L-1 when using nitrate and 200 mg COD L-1 when
using nitrite. Upon these additions, denitrifying
bacteria will become active and their activity will
usually tend to increase the pH. The automatic
titration system will react to decrease the pH to
maintain the pH set point through acid addition.
3. Record the volume of titration solution added over
time (Vtit versus time). Check that the pH value
remains within the interval pH set point ± 0.02. DO
concentration should be below the detection limit.
Continue the test until a clear bending point is
observed in the cumulated titration solution plot. This
bending point indicates that the nitrate is fully
depleted and consequently denitrification has
stopped. The collected data should be sufficient to
reliably estimate the titration rate (Qtit) from a linear
regression of Vtit versus time with a satisfactory
coefficient of determination (normally R2 > 0.98).
4. Step 3 can be repeated by adding another dose of
nitrate. The addition of SB is no longer needed since
a sufficient residual concentration is still present in
the activated sludge to support a second
denitrification phase.
5. End the test according to the instructions reported in
Step 4 of the test operation procedure described for
Test NIT.TIT.1.
NO3
Qtit
Vtit
YNO3_SB,Ax =
Data analysis
A typical trend in the cumulated volume of titration
solution added during the test is depicted in Figure 2.4.9.
Qtit
RBCOD
NO3
Time
Figure 2.4.9 Example of a pH-static titration curve during a denitrification
test to assess the maximum denitrification rate. Arrows indicate the
addition of nitrate and SB solutions. The relevant titration rate and is also
shown in the graph.
From the data (Vtit versus time data), the titration
rates (Qtit) can be computed based on the slope of the
titration curve (see Section 2.4.3).
The YNO3_H+,Ax, in g N mol Protons-1, can be assessed
by considering the volume of titration solution added
until the plateau of the titration curve is observed (VT in
Figure 2.4.9) and the mass of nitrate (or nitrite) added
(MNOx,ini):
YNO3_H+,Ax =
MNOx,ini
VT · NT
Eq. 2.4.55
Where, NT is the titration solution normality. When
more than one spike is performed (as in the case shown
in Figure 2.4.9), the calculation can be repeated per each
spike and the mean value can be taken as the estimate of
YNO3_H+,Ax.
The titration rate (Qtit in mL min-1) can be used to
assess the denitrification rate (FNO3_H+,Ax in mg N min-1)
by taking into account the concentration of the titration
solution (NT) and the YNO3_H+,Ax value:
FNO3_H+,Ax = Qtit · NT · YNO3_H+,Ax
Eq. 2.4.56
ACTIVATED SLUDGE ACTIVITY TESTS
99
Finally, the maximum specific denitrification rate of
the sludge (qNOx_N2, in mg N g VSS-1 h-1), is computed by
taking into account the MLVSS concentration of the
sludge sample (in g VSS L-1) and the suspension volume
assessed at the end of the test (in L):
qNOx_N2 = 60 · FNO3_H+,Ax / (VML · XVSS )
Eq. 2.4.57
2.4.8 Anammox batch activity tests:
preparation
As described in Section 2.4.1, the anammox process
consumes ammonium and nitrite and converts them
mainly into nitrogen gas as well as a minor fraction of
nitrate. When the anammox process takes place in a batch
reactor then, variations in the ammonium, nitrite and
nitrate concentrations are expected and can be monitored
to follow the time evolution of the process. Furthermore,
when a gas-tight reaction vessel is used to perform the
batch test, the release of dinitrogen gas will cause a
pressure increase, which can also be monitored over time
to measure the reaction kinetics. Therefore, two
alternatives are available to track the evolution of the
anammox process:
• Chemical
tracking,
by
assessing
the
ammonium/nitrite/nitrate concentration evolution in
time.
• Manometric tracking by assessing the overpressure
caused by dinitrogen release in a gas-tight reactor.
Each one of these alternatives is discussed hereafter
in the coming two tests. The first test (AMX.CHE) refers
to a batch test performed by chemical tracking to assess
the maximum activity of an anammox culture fed with
synthetic autotrophic medium. In this test, the
stoichiometric coefficients NO2/NH4 and NO3/NH4 ratios
will be also assessed. In the second test (AMX.MAN),
anammox biomass is suspended in a real wastewater and
a manometric tracking procedure is applied to evaluate
its treatability via the anammox process. In this example,
the inhibition potential of a wastewater is evaluated by
comparing the maximum rate obtained in the presence of
wastewater with the maximum rate observed in the
presence of a synthetic medium.
2.4.8.1 Apparatus
Each tracking methodology (chemical or manometric)
has special apparatus requirements.
When chemical tracking is applied, refer to the
following list of equipment:
1. An (airtight) batch reactor equipped with a mixing
system and adequate sampling ports (as described in
Section 2.4.3.1).
2. A nitrogen gas supply (recommended).
3. A
calibrated
pH
electrode
(if
not
included/incorporated in the batch reactor setup).
4. A 2-way pH controller via HCl and NaOH addition
(alternatively a one-way control - generally for HCl
addition - or manual pH control can be applied
through the manual addition of HCl and NaOH).
5. A thermometer (recommended working temperature
range of 0 to 40 °C).
6. A temperature control system (if not included in the
batch reactor setup).
7. A
DO
meter
with
electrode
(if
not
included/incorporated in the batch reactor setup) to
verify anoxic conditions.
8. A stopwatch.
When the manometric tests are performed, refer to
the following list of equipment:
1. The equipment given for the manometric system,
described in Section 2.4.3.3.
2. A nitrogen gas supply (recommended).
2.4.8.2 Materials
For a complete list of required materials refer to Section
2.2.3.2 (with the exception of points 7 and 8).
2.4.8.3 Working solutions
• Real wastewater
For batch tests that require the use of a real wastewater,
follow the instructions given in Section 2.2.3.3.
• Synthetic medium
If tests can be or are desired to be performed with
synthetic wastewater, the synthetic influent media could
contain a mixture of ammonium, nitrite and bicarbonate
plus necessary (macro- and micro-) nutrients. Generally,
they can be mixed all together in the same media or
prepared separately if they need to be added in different
phases or times. The usual compositions and
concentrations are (based on van de Graaf et al., 1996): 1
g L-1 (NH4)2SO4 (7.6 mM, 106 mg N L-1), 0.25 g L-1
NaNO2 (3.6 mM, 51 mg N L-1), 0.6 g L-1 NaNO3 (7.1
mM, 99 mg N L-1), 1 g L-1 NaHCO3 (11.9 mM), 0.025 g
L-1 KH2PO4 (0.18 mM), 0.1 g L-1 MgSO4·7H2O (0.41
mM), 0.15 g L-1 CaCl2·2H2O (1.02 mM) and 1.25 mL L-1
trace elements solutions A and B (see below for trace
elements A and B preparation). Similar nutrient solutions
100
can be used as long as they contain all the previously
reported required nutrients. The trace element solutions
should contain all the required micro-nutrients (iron,
zinc, cobalt, manganese, copper, molybdate, nickel,
selenium and boron) to ensure that cells are not limited
by their absence and avoid obtaining incorrect results and
in extreme cases the failure of the test. Thus, despite the
fact that their concentrations may seem very low, one
must make sure that all of the constituents are added to
the solution in the required amounts. The following
composition (amounts per litre of micro-nutrient
solution) is recommended (based on van de Graaf et al.,
1996): Trace elements solution A: 5 g EDTA, 9.14 g
FeSO4·7H2O, trace element solution B: 15 g EDTA, 0.43
g ZnSO4·7H2O, 0.24 g CoCl2·6H2O, 0.99 g MnCl2·4H2O,
0.25 g CuSO4·5H2O, 0.22 g Na2MoO4·2H2O, 0.19 g
NiCl2·6H2O, 0.21 g NaSeO4·10H2O, and 0.014 g H3BO3.
• Ammonium or nitrite solutions
For the dosage of the anammox substrates, ammonium
and nitrite stock solutions can be prepared (e.g. 1-10 g N
L-1) using either (NH4)2SO4 or NH4Cl salts for
ammonium and NaNO2 or KNO2 salts for nitrite.
Generally, they can be mixed together in the same stock
solution or prepared separately if they need to be added
in different phases or at different times. When an
ammonium/nitrite solution is prepared the molar ratio
between the two substrates is usually set to one in order
to ensure an excess of ammonium throughout the test.
Bicarbonate can also be added to the batch test to avoid
inorganic carbon (IC) limitation: in order to ensure IC in
excess during the test, add bicarbonate to the ammonium
solution up to a molar ratio equal to 0.7 mol-IC molNH4-1 (i.e. ten times the stoichiometric requirements of
0.066 mol-IC mol-NH4). The latter should be considered
in case the wastewater tested has limiting concentrations
of inorganic carbon and/or if the oxygen is removed via
sparging CO2-free gases (e.g. N2), which will lead to CO2
stripping.
• Nitrate solution
In order to ensure the correct redox conditions, nitrate can
be dosed prior to the execution of the batch test using a
nitrate stock solution that can be prepared (e.g. 1-10 g N
L-1) using either NaNO3 or KNO3 salts. This is especially
needed when consecutive batch tests need to be
performed using the same anammox sludge (e.g. to
evaluate the long-term effect of exposure to a particular
compound) in order to avoid sulphate reduction to
sulphide (H2S), which may be toxic to anammox bacteria.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
• Washing media
If the sludge sample must be 'washed' to remove any
undesirable compounds (which may even be inhibitory
or toxic), a washing solution is needed. The synthetic
medium described above can be used as a washing
medium by simply removing the nitrite salts from the
recipe. The washing process can be repeated twice or
three times. Thereafter, the following preparation steps
can be performed. In special cases, e.g. when sludge from
a full-scale plant is used, the plant effluent may be used
for washing purposes (assuming that its composition
allows this).
• Acid and base solutions
These should be respectively 100-250 mL of 0.2 M HCl
and 100-250 mL 0.2 M NaOH solutions for automatic or
manual pH control, and 10-50 mL of 1 M HCl and 10-50
mL 1 M NaOH solutions for initial pH adjustment if the
desired operational pH is very different from the pKa
value of the buffering agent.
Finally, the required working and stock solutions to
carry out the determination of the analytical parameters
of interest must also be prepared in accordance with
Standard Methods (APHA et al., 2012) and the
corresponding protocols.
It is recommended to take samples of the media and
activated sludge prior to the execution of the experiment
to confirm/check the initial (desired) concentrations of
parameter(s) of interest (e.g. ammonium, nitrite, nitrate,
XTSS and XVSS ).
2.4.8.4 Material preparation
For general instructions on how to organize the material
preparation, refer to Section 2.2.3.4 and to Table 2.2.2. If
required, test-specific requirements will be listed within
each test protocol.
2.4.9 Anammox batch activity tests:
execution
Test AMX.CHE Anammox chemical test: assessing the
maximum anammox kinetic rate and the stoichiometric
coefficient NO2/NH4 and NO3/NH4 ratios
Activated sludge preparation
For a suitable sampling point for the activated sludge,
refer to Section 2.4.3.4.
1. For conventional activated sludge samples from
municipal wastewater treatment plants, a sample
ACTIVATED SLUDGE ACTIVITY TESTS
5.
Test execution
1. Verify that the temperature, pH and DO target values
are within the desired intervals. Otherwise adjust
them and wait until they stabilize.
2. Add the ammonium/nitrite stock solution (previously
adjusted to the target temperature) to achieve a
neither limiting nor inhibiting nitrite concentration in
the activated sludge: 50 to 75 mg N L-1 is usually
adequate for anammox biomass cultivated under nonstrict nitrite limiting conditions (see Section 2.4.3.6
for further explanations). If any residual nitrite is
present in the activated sludge sample, the nitrite
addition should be reduced accordingly. The
ammonium starting concentration is less critical
considering the concentration ranges typical of
anammox systems due to the lower inhibiting effect
4.
5.
6.
Note that the nitrite concentrations are unstable, and
therefore they need to be quickly determined after
sampling.
Data analysis
A typical output of this test is shown in Figure 2.4.10.
NO2
,S
NO3
(mg N L-1)
80
,S
4.
3.
NH4
3.
on anammox bacteria: 50 to 200 mg N L-1 is usually
considered as adequate.
Start the stopwatch just after the addition of the
ammonium/nitrite to track the sampling times.
Collect the activated sludge samples at regular time
intervals. For instance, collect the samples every 20
to 30 min throughout the duration of the test (which
usually lasts around 3 to 4 h).
Conclude the test when the nitrite is fully depleted.
Use nitrite strip tests to quickly estimate the nitrite
concentrations.
Take a sample to determine the final MLVSS
concentration.
70
60
rAMX,NH4_NO3
50
40
rAMX,NH4
30
rAMX,NO2
20
10
S
2.
containing a XVSS of around 2-10 g L-1 will be
suitable. The XVSS can be adjusted as suggested in
Section 2.4.3.5. For anammox granular sludge,
instead of using the activated sludge, granular
biomass can be easily separated from the supernatant
by settling (in a cylinder or in an Imhoff cone) and resuspended in a washing solution or in the effluent of
the plant. Preliminary tests can be conducted to
evaluate the density of the settled granular sludge (g
VSS L-1) to pour a defined amount of anammox
granular sludge (g VSS) into the reaction vessel.
Pour a defined volume of the activated sludge sample
(typically 1 to 3 L) or a known amount of anammox
granular sludge (re-suspended in the medium to be
tested) into the reaction vessel and start mixing. Also,
start the temperature and pH control systems to
maintain both parameters around the desired set point
values.
Sparge N2 (or N2/CO2 gas mixture, see Section
2.4.3.5) into the headspace for approximately 10 min
to ensure an oxygen-free environment. Install a gas
outlet to limit the overpressure. Gas sparging can
continue until the end of the test. If this is not feasible,
then an airtight reactor with a device able to prevent
oxygen intrusion can be used (e.g. using a
unidirectional check valve or a water lock) (see
Section 2.2.2.1 for further details).
Dose nitrate up to a final concentration in the
activated sludge of 50-100 mg N L-1 in order to ensure
an adequate redox potential and avoid sulphate
reduction.
Wait for approximately 30 min to ensure stable initial
conditions. This pre-incubation phase will normally
allow the removal of any residual nitrite present.
101
0
0
0.5
1.0
Time (h)
1.5
2.0
Figure 2.4.10 Typical N profiles in a Test AMX.CHE: ammonium ( ● ),
nitrite (●) and nitrate (●) concentrations are displayed on the y-axis. The
relevant rates of interest are also indicated (ammonium removal rate
rAMX,NH4, nitrite removal rate rAMX,NO2, and nitrate production rate rAMX,NH4_NO3).
The linear regression over these data (see Section
2.4.5) can be used to determine the ammonium (rAMX,NH4)
and nitrite removal rates (rAMX,NO2) as well as the nitrate
production (rAMX,NH4_NO3) rate (expressed as mg N L-1 h-1).
The anammox rate is usually expressed as the dinitrogen
gas produced, which is equivalent to the nitrogen
removed from the wastewater. The maximum specific
anammox rate (qAMX,N2 as mg N2-N g VSS-1 h-1) can
therefore be computed as follows:
102
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
qAMX,N2 =
rAMX,NH4 + rAMX,NO2 ‒ rAMX,NH4_NO3
Eq. 2.4.58
XVSS
In the previous expression, when anammox granular
sludge is used to perform the test, then the biomass
concentration at the denominator (the term XVSS) will
correspond to the g VSS added to the reactor vessel
during the test preparation divided by the mixed liquor
volume.
The stoichiometric coefficient ratios of interest, such
as YNH4_NO2,AMX and YNH4_NO3,AMX, can be easily
calculated from the relative removal/production rates
using the following expressions:
Y
_
,
=
rAMX,NO2
rAMX,NH4
Eq. 2.4.59
Y
_
,
=
rAMX,NH4_NO3
rAMX,NH4
Eq. 2.4.60
Test AMX.MAN Anammox manometric test: assessing the
maximum anammox kinetics
Activated sludge preparation
1. Follow step 1 of the activated sludge preparation
protocol described in Test AMX.CHE above.
2. Select an appropriate activated sludge volume to be
poured into the reaction vessel (VML). For anammox
granular sludge, see the instructions described in Test
AMX.CHE. In this test, the overpressure depends on
the ratio between the amount of ammonium to be
oxidized (which is based on the stoichiometry of the
reaction; 1 mol of NH4 consumption will lead to 1
mol of N2 production) (see Section 2.4.1) and the
headspace volume (VHS in L). Hence, the correct
selection of this ratio is crucial to avoid the
generation of an extreme overpressure (either too low
or too high). The maximum overpressure will be
achieved at the end of the test, i.e. when nitrite
(usually the limiting substrate) has been fully
converted. Considering a YNH4_NO2,AMX stoichiometric
ratio of 1.32 mol-NO2 mol-NH4-1, the maximum
overpressure (Pmax) can be estimated as follows:
Pmax ‒ Patm =
Patm MNO2_N2
(273 + TC )
·
· 22.4 ·
VHS
14
273
Eq. 2.4.61
Where, MNO2_N2 (in mg N) is the mass of nitrite
that is converted during the course of the test,
dependent on the initial nitrite concentration (SNO2,ini,
in mg N L-1) and activated sludge volume (VML in L):
MNO2_N2 = SNO2,ini · VML
Eq. 2.4.62
By substituting this equation into the previous
one and rearranging, the following relationship is
obtained:
273
1,000
VML Pmax ‒ Patm 14 · 1.32
=
·
·
·
VHS
Patm
22.4
(273 + T ) SNO2,ini
Eq. 2.4.63
The optimal Pmax value usually lies around 0.2
atm, which means that, under typical test conditions
(TC ~30 °C, SNO2,ini ~50 mg N L-1), a VML to VHS ratio
of around 3 is obtained. Therefore, for a total reactor
volume of 1 L (VTOT = VML + VHS), an adequate
activated sludge volume will be 0.75 L.
3. Pour the previously estimated activated sludge
volume into the reaction vessel. Place the reaction
vessel on the shaker (see Section 2.4.3.1 for mixing
requirements). Flush the reactor headspace with N2
(or a N2/CO2 gas mixture, see Section 2.4.3.3) and
seal the gas bottle tightly. Place the reactor into the
thermostatic chamber, apply a gentle mixing and wait
for 30 min until the temperature stabilizes.
Test execution
1. Select the volume of nitrite solution to be dosed at the
beginning of the batch test to achieve a neither
limiting nor inhibiting nitrite concentration in the
activated sludge: typically, 50 to 75 mg N L-1 are
adequate concentrations for anammox biomass
cultivated under non-strict nitrite limiting conditions
(see Section 2.4.3.4 for further details). Usually, the
initial ammonium concentration is similar to the
nitrite concentration. Higher initial ammonium
concentrations can be used, as long as they remain
below the reported inhibitory level for anammox
bacteria (< 1 g N L-1). If there is residual nitrite
present in the activated sludge sample, the nitrite
addition should be reduced accordingly. Using a
syringe, inject the ammonium/nitrite stock solution
through the rubber septum.
2. Start the manometric data collection. Collect the data
every 30 to 60 min until there is a point where the
overpressure curve bends, which will correspond to
the exhaustion of nitrite.
3. End of the test. Check the final pH and take a final
sample to determine the MLVSS concentration.
ACTIVATED SLUDGE ACTIVITY TESTS
103
Data analysis
Using the ideal gas law, the pressure data can be
converted into the gas moles (N2 in this case) emitted to
the headspace:
n(t) =
P(t) · VHS
R · TK
Eq. 2.4.64
Where, n(t) is the number of N2 moles present in the
headspace volume (VHS) at time t, P(t) is the pressure in
the headspace at time t, R is the ideal gas constant and TK
the temperature of execution expressed in Kelvin.
Figure 2.4.11 shows a typical profile of a manometric
anammox test executed with a manometer equipped with
a data logger after the conversion of the recorded
pressure data into moles of N2 gas produced. Disregard
the data collected in the first 10-15 min of the test since
various phenomena may overlap affecting the headspace
overpressure, such as: (i) residual oxygen intrusion from
the headspace, (ii) water-vapour pressure equilibrium;
(iii) initial N2 accumulation in the liquid phase, and (iv)
microbiological lag phases.
0.7
N2 produced (mmol)
0.6
0.5
0.4
FAMX,NH4_N2
0.3
+
NH4
NO2-
0.2
0.1
0
0
100
200
300
400
Time (min)
500
600
700
Figure 2.4.11 Typical N2 gas profiles obtained in a Test AMX.MAN
performed with a manometer equipped with a data logger. The maximum
N2 production rate FAMX,NH4_N2 is depicted in the figure.
The cumulative N2 production curve can be used to
compute the N2 production rate (FAMX,NH4_N2, N2-mol
min-1) using linear regression (see Section 2.4.5).
Moreover, data can be further expressed in mg N2-N min1
using the equivalent molecular weight of dinitrogen gas
(28 g N N2-mol-1). The specific anammox rate qAMX,N2
(mg N g VSS-1 h-1) can be computed using the MLVSS
concentration:
qAMX,N2 = 60 · FAMX,NH4_N2 / (VML · XVSS )
Eq. 2.4.65
2.4.10 Examples
2.4.10.1 Nitrification batch activity test
Description
To illustrate the execution of an aerobic batch activity
test for nitrification, data from a test performed at 20oC
with a full-scale activated sludge sample is presented in
this section. The Test NIT.CHE.1 was carried out to
determine the maximum specific biomass ammonium
oxidation rate. The batch activity test was performed
using a 2.5 L reactor. All the equipment, apparatus and
materials were prepared as described in Section 2.4.4.1.
The pH and DO sensors were calibrated less than 24 h
before the test execution. The batch test lasted 3 h.
Including the time needed for the test preparation and the
cleaning phase afterwards, the operations lasted
approximately 5 h. Prior to the batch test, 2.0 L of fresh
activated sludge collected at the end of the aerobic phase
of a full-scale plant (according to Section 2.4.3.4) was
transferred to the reactor and acclimatized for 1 h at 20
°
C under slow mixing (100 rpm) following the
recommendations described in Section 2.4.3.5. The pH
and DO control were started up at pH and DO set points
of 7.5 and 6 mg O2 L-1, respectively. Thereafter, 15 min
before the start of the test, mixing was increased to 200
rpm and samples were collected for the determination of
the parameters of interest (e.g. ammonium, nitrite, nitrate
and MLVSS concentrations). Before the addition of the
synthetic medium, its temperature was adjusted in a
water bath to the target temperature of the batch test (20
º
C). The first sample was collected 5 minutes before the
addition of the synthetic medium to determine the initial
conditions. The test started with the addition of the
synthetic media (minute zero): 50 mL containing 1 g
NH4-N L-1 as well as other macro- and micro-nutrients as
described in Section 2.4.3.6. Samples were collected
every 30 min over a period of 3 h. The duration of the
batch test was chosen in order to allow a complete
oxidization of the ammonium present. Immediately after
collection, all the samples were prepared and preserved
as described in Section 2.2.3.4. All the collected samples
were analysed as described in Section 2.4.3.7. Table 2.4.5
shows the experimental implementation plan and the
results of the execution of the test.
Data analysis
Following on from the results from the experiment shown
in Table 2.4.5, Figure 2.4.12 shows the results from the
test and also an estimation of the maximum volumetric
kinetic rates (by applying linear regression).
104
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 2.4.5 Results from the nitrification batch activity test.
Nitrification batch activity test
Code: NIT.CHE.1
Date:
Description:
Test No.:
Duration
Monday
05.10.2015 10:00 h
Nitrification test at 20 °C with real activated sludge
1
3 h (180 min)
Substrate:
Sampling point:
Samples No.:
Total sample volume:
Synthetic: ammonium and nitrite (1,000 mg L ) + minerals
Middle mixed liquor height in the SBR
NIT.CHE 1-8
80 mL
(5 mL normal sample, 20 mL for MLVSS)
2.5 L
-1
Reactor volume:
Sampling schedule
Time (min)
Time (h)
Sample No.
Parameter
-15
-0.25
1
-1
6.8
-1
0
NH4-N (mg N L )
NO 2-N (mg N L )
1
NO 3-N (mg N L )
-5
-0.08
1
0.1
-1
MLSS and MLVSS (mg L )
1
Experimental procedure in short:
1. Confirm availability of sampling material and required equipment.
2. Confirm calibration and functionality of systems, meters and sensors.
3. Transfer 2.0 L of sludge to batch reactor.
0
0.00
2
4. Start aerobic conditions with gentle mixing and air sparging at T and pH set points.
5. 15 min before starting, take sample for initial conditions (sample NIT.CHE 1).
7. Start batch test: add 0.05 L of synthetic influent.
8. Take first sample for the determination of the initial conditions (minute zero)
9. Minute 30, continue sampling program according to schedule
10. Minute 180, stop aeration and mixing.
11. Organize the samples and clean the system.
12. Verify that all systems are swtiched off.
30
60
0.50
1.00
3
4
AEROBIC PHASE
90
1.50
5
120
2.00
6
150
2.50
7
31
27.3
20.6
17.2
12.4
4.9
0.4
0
0.4
0.9
0.6
0.3
0.4
0.1
0.1
5.1
9.8
14.2
19.4
24
See table
08:40
09:40
09:55
10:00
10:30
13:00
13:10
13:20
180
3.00
8
29.6
See table
Average value of the concentration present in the synthetic substrate and in the liquid phase of the sludge sample prior the start of the test
MLSS & MLVSS measurements
Sampling point
Cup No.
W1
W2
W3
W2-W1
W2-W3
MLSS
MLVSS
Ratio
Start test
1
2
3
0.09630
0.09580
0.09640
0.16530
0.16380
0.16440
0.10210
0.10190
0.10230
0.06900
0.06800
0.06800
End test
4
5
6
0.09540
0.09610
0.09570
0.16490
0.16410
0.16400
0.10200
0.10180
0.10220
0.06950
0.06800
0.06830
0.06320
0.06190
0.06210
Average
0.06290
0.06230
0.06180
Average
3,450
3,400
3,400
3,417
3,475
3,400
3,415
3,430
3,160
3,095
3,105
3,120
3,145
3,115
3,090
3,117
0.92
0.91
0.91
0.91
0.91
0.92
0.90
0.91
2
Concentrations corrected considering the dilution due to the synthetic medium addition.
297
313
-1
Ash (mg L )
The maximum volumetric rates for ammonium
removal and nitrate production are 10.3 and 9.7 mg N L1
h-1, respectively. Taking into account the average
MLVSS concentration of 3.1 g L-1 between the sludge
samples collected at the beginning and end of the test, a
biomass-specific ammonium oxidation rate of 3.3 mg N
g VSS-1 h-1 can be estimated, which corresponds to about
80 mg N g VSS-1 d-1. It is important to note that the sum
of soluble inorganic nitrogen compounds at the
beginning and end of the batch test are comparable (31.1
vs 30.1 mg N L-1), indicating that nitrification was the
dominant process during the batch test.
NO3
3,117
0.91
,S
3,430
3,120
0.91
MLVSS (mg L )
Ratio
(mg N L-1)
End test
3,417
NO2
Start test
-1
,S
-1
MLSS (mg L )
NH4
Biomass composition
Sampling point
35
y = -10.343x + 31.771
2
R = 0.9927
30
y = 9.707x + 0.039
2
R = 0.9992
25
20
15
10
5
S
2
Time (h:min)
08:00
08:10
08:30
0
0
0.5
1.0
1.5
Time (h)
2.0
2.5
3.0
Figure 2.4.12 Graphic representation of the ammonium (●), nitrite (●)
and nitrate concentrations ( ● ) obtained in the example of the batch
activity nitrification test (Test NIT.CHE.1). The test was performed with a
full-scale activated sludge at 20 ºC and pH 7.5 using a synthetic medium as
influent.
ACTIVATED SLUDGE ACTIVITY TESTS
105
The two trend lines show the ammonium conversion
and nitrate production rates for further estimation of the
corresponding maximum specific biomass kinetic rates.
A decrease in the sum of ammonium, nitrite and nitrate
concentrations would suggest the simultaneous
occurrence of nitrogen removal processes such as
denitrification or anammox, which can take place in the
presence of anoxic conditions. The comparable
ammonium oxidation and nitrate production rates
indicate that, in the tested sludge, the activities of AOO
and NOO are well-balanced and allowed to achieve full
nitrification. These observations were also supported by
the absence of nitrite. The measured specific ammonium
oxidation rate is comparable to the kinetic rates reported
in Table 2.4.1, indicating a high enrichment of nitrifying
bacteria in the activated sludge sample tested.
2.4.10.2 Denitrification batch activity test
Description
In this section, a test to assess the maximum
denitrification rate and the anoxic growth yield in the
presence of acetate as the carbon source is described.
An airtight reactor equipped with a mechanical
mixing system and automatic pH and temperature control
systems was used. N2 gas bubbling was also provided.
The day before testing, the pH and DO probes were
calibrated and all the required materials were prepared as
suggested in Section 2.2.3.2 (with the exception of points
7 and 8). A 10 g N L-1 nitrate solution was prepared with
NaNO3 and an acetate stock solution using sodium
acetate at a concentration of 10 g COD L-1 (taking into
account that 1 g CH3COONa corresponds to 0.78 g
COD). 1 M HCL and 1 M NaOH solutions were prepared
for automatic pH correction. Moreover, 2 L of sludge
were used. A preliminary working plan was defined as
displayed in Table 2.4.6.
Table 2.4.6 Results from the denitrification batch activity test.
Denitrification batch activity test
Date:
Description:
Test No.:
Duration
Substrate:
Sampling point:
Samples No.:
Total sample volume:
Reactor volume:
Code: DEN.CHE.1
Wednesday 02.09.2015 9:00 h
Tests at 20 °C, pH 7, artificial substrate & enriched lab culture
1
3.5 h (210 min)
Synthetic: Acetate (200 mgCOD/L) + nitrate (20 mgN/L)
Middle mixed liquor height in the SBR
DEN.CHE.1 (1-18)
240 mL
(10 mL for MLVSS, and 10 mL each sample)
2.5 L
Sampling schedule
Time (min)
Time (h)
Sample No.
Parameter
Experimental procedure in short:
Time (h:min)
1. Day before instrumentation check, probes calibration preparation of the
working plan, stock solutions, containes for sample collections, and all other materials.
2. Sludge sampled from SBR and transfered to reaction vessel.
09:00
3. Activation of N2 sparging.
09:05
4. Nitrate addition, fist sample taken.
09:10
5. Other 5 samples taken at 20 min intervals.
6. Acetate addition and sampling.
10:30
7. Other 13 samples taken at 5/10 min intervals.
8. Test stopped.
11:35
9. Sample for MLVSS measurement taken, total volume assessment.
11:40
10. Verify that all samples are correctly stored and switch off system.
11:50
0
0.00
1
20
0.33
2
40
0.67
3
60
1.00
4
80
1.33
5
-1
23.0
23.2
22.0
21.8
20.2
18.2
-1
0.1
0.0
0.2
0.0
0.0
1.0
418.0
407.3
180
3.00
17
185
3.08
18
NO3 -N (mg N L )
NO2 -N (mg N L )
-1
CODsoluble (mg COD L )
Sampling schedule (continued)
Time (min)
Time (hrs)
Sample No.
Parameter
130
2.17
13
140
2.33
14
-1
4.9
1.8
0.3
0.0
-1
2.0
2.1
1.5
1.0
0.0
315.0
303.6
285.7
275.0
271.4
NO3 -N (mg N L )
NO2 -N (mg N L )
-1
CODsoluble (mg COD L )
-1
MLSS (g L )
150
160
2.50
2.67
15
16
ANOXIC PHASE
85
90
1.42
1.50
6
7
ANOXIC PHASE
0.0
2.51
Total biomass (gVSS)
-1
4.58
-1
r D,end (mg N L min )
-1
0.035
-1
r D,exog (mg N L min )
-1
0.278
-1
2.050
r COD (mg COD L min )
0.660
Anoxic growth yield
-1
-1
Denitrification rate (mg N g MLVSS h )
6.600
95
1.58
8
100
1.67
9
105
1.75
10
110
1.83
11
120
2.00
12
15.8
14.0
12.3
11.2
10.2
7.3
2.0
2.5
2.8
3.0
3.0
2.7
397.3
385.8
378.0
366.5
353.7
336.5
106
Data analysis
Relevant implementation data and results of the
analytical determinations are reported in Table 2.4.6. In
Figure 2.4.13, the COD and SNO3,Eq (computed as SNO3,Eq
= SNO3 + 0.6 · SNO2) trends are plotted.
As shown in Figure 2.4.13, the endogenous and
exogenous denitrification rates were estimated by linear
regression (rNOx_N2,exo and rNOx_N2,endo in mg N L-1 min-1),
as well as the exogenous COD consumption rate (rCOD, in
mg COD L-1 min-1). Note that only data obtained in the
non-limiting denitrification phase were used for the
determination of the kinetic rates. For example, only
those data that led to the highest linear fitting, based on
the highest R2 value, for both N-NOeq and soluble COD.
450
y = -0.035x + 23.44
2
R = 0.8578
20
y = -2.058x + 582.34
2
R = 0.9981
400
350
15
300
10
y = -0.278x + 42.203
R2= 0.9945
5
SAc,COD (mg COD L-1)
25
S NO3,eq (mg N L-1)
The initial nitrate concentration was set to 20 mg NNO3 L-1 (corresponding to the addition of 4 mL of nitrate
stock solution). The initial COD concentration was set to
200 mg L-1 (an addition of 40 mL of acetate stock
solution). According to the instructions reported in
Section 2.4.7, up to 18 samples were planned to be
collected. All corresponding materials and consumables
(e.g. plastic cups) were prepared and labelled to avoid
identification errors during the sample collection. On the
day of the test, the activated sludge was sampled from a
bench scale SBR at the end of its denitrification phase.
Two litres of activated sludge were transferred to the
reaction vessel and the temperature set point was set at
20 °C, similar to the operating temperature of the benchscale SBR. Simultaneously, the pH-control system was
started up with a set point value of 7.4. The initial pH was
7.7. The N2 gas sparging system was turned on to remove
any residual DO and ensure anoxic conditions after the
nitrate addition. After 5 min, the SO2 concentration in the
bulk liquid was below the probe detection limit and the
N2 sparging was turned off. The predefined volume of the
nitrate stock solution was spiked through the sampling
port. After approximately 1 min, the stopwatch was
turned on and the first sample was taken (minute zero).
Four more samples were taken at 20 min intervals to
assess the endogenous denitrification rate. Thereafter, the
COD stock solution was added and, 30 seconds
afterwards, a sample was taken. Later on, 6 samples were
taken at 5 min intervals and afterwards samples were
taken every 10 min. The sampling campaign continued
while the nitrate and nitrite profiles were continuously
followed using NO3 and NO2 detection paper strips until
the nitrate and nitrite were fully consumed. These
samples were used to assess the nitrate, nitrite, and
soluble COD concentration. The test finished after 180
min. One final sample was taken to assess the MLVSS
concentration. Eventually, all the systems were stopped
and the reactor was opened in order to measure the final
volume of activated sludge. All the samples were
collected and preserved as described in Section 2.2.3.4.
All the collected samples were analysed as described in
Section 2.4.3.7.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
250
0
0
50
100
Time (min)
150
200
200
Figure 2.4.13 Graphic representation of (●) the N-NOeq and (●) soluble
COD concentrations obtained in the Test DEN.CHE.1 (as presented in Table
2.4.6). The test was performed using an activated sludge sample from a
SBR operated at 20 ºC and pH 7.5 using acetate as the carbon source. The
slopes of the trend lines were used to quantify the endogenous, exogenous
denitrification and organic carbon uptake rates.
Thus, the maximum rates can be determined as:
•
q
Maximum specific denitrification rate:
_
60 ·
•
= 60 ·
rNOx_N2,exo ‒ rNOx_N2,endo
X
0.28 ‒ 0.035
= 5.8 mg N g VSS-1 h-1
2.51
Eq. 2.4.66
Anoxic growth yield on acetate:
YOHO,Ax = 1 ‒ 2.86 ·
1 ‒ 2.86 ·
rNOx_N2,exo ‒ rNOx_N2,endo
rCOD
0.28 ‒ 0.035
2.05
= 0.66 g CODbiomass g CODacetate -1
Eq. 2.4.67
Both obtained values are in the range of data reported
in Table 2.4.2, supporting the reliability of the results.
ACTIVATED SLUDGE ACTIVITY TESTS
The maximum activity should be evaluated under
non-limiting conditions. Considering the half-saturation
constant for nitrite of 0.035 mg N L-1 reported in
literature (in most cases the limiting substrate) (Lotti et
al., 2014), non-limiting nitrite concentrations are in the
Data analysis
Following the results from the experiment shown in
Table 2.4.7, Figure 2.4.14 shows the results from the test
displayed in the implementation plan and also an
estimation of the maximum volumetric kinetic rates.
NO2
,S
NO3
(mg N L-1)
80
,S
To illustrate the execution of an anammox batch activity
test, data from a test performed at 30 ºC with a full-scale
anammox biomass sample is presented in this section.
The Test AMX.CHE was carried out to determine the
maximum specific biomass anammox rate (kAMX, mg N g
MLVSS-1 d-1) as well as the stoichiometric parameters of
interest for nitrite consumption and nitrate production
with respect to ammonium consumption (NO2/NH4 and
NO3/NH4 ratios described in Section 2.4.3.8). The batch
activity test was performed using a 2.5 L reactor. All the
equipment, apparatus and materials were prepared as
described in Section 2.4.3. The pH sensor was calibrated
less than 24 h before the test execution. The batch test
lasted for 3 h. Including the time needed for the test
preparation and the cleaning phase afterwards, the
operations lasted for approximately 5 h. Prior to the batch
test, 1 L of fresh anammox sludge collected at a full-scale
1-stage PN/anammox plant (according to Section 2.4.3.4)
was transferred to the reactor and acclimatized for 1 h at
30 ºC under slow mixing (100 rpm) following the
recommendations described in Section 2.4.9. To ensure
an oxygen-free environment, a N2/CO2 gas mixture (see
Section 2.4.3.5) was sparged into the bulk liquid in the
first 10 min and into the reaction headspace throughout
the batch test. A gas outlet was provided to limit any
overpressure and prevent oxygen back-diffusion (e.g.
using a water lock). The pH control was started with a pH
set point of 7.5. Thereafter, 15 min before the start of the
test, mixing was increased to 200 rpm and samples were
collected to characterize the anammox sludge used in the
test (e.g. ammonium, nitrite, nitrate and MLVSS
concentrations). The first sample was collected 5 min
before the addition of a synthetic medium in order to
determine the initial conditions. The test started with the
addition of 70 mL of the synthetic medium containing 1
g NH4-N L-1 and 1 g NO2-N L-1 as well as other macroand micro-nutrients as described in Section 2.4.9 (minute
zero). In the example, nitrate was not dosed at the
beginning of the test since it was already present in the
anammox sludge sample. Before its addition, the
temperature of the synthetic medium was adjusted in a
water bath to the target temperature of the test (30 °C).
Samples were collected every 30 min for a period of 3 h.
NH4
Description
order of 1-2 mg N L-1 for flocculent or suspended
anammox biomass. For biomass types characterized by
higher mass transfer limitations such as biofilms attached
to inert carriers and granular sludge, non-limiting
conditions can be ensured at a nitrite concentration in the
order of 5-10 mg N L-1, depending on the density of the
biofilm and on the specific anammox activity of the
biofilms. When detailed information on the mass transfer
characteristics of the biomass used during the batch test
is not well known, one should consider for data analysis
only those concentrations that can be satisfactory
interpolated by linear regression (R2 > 0.95). Immediately
after collection, all the samples were prepared and
preserved as described in Section 2.2.3.4. All the
collected samples were analysed as described in Section
2.4.3.7. Table 2.4.7 shows the experimental
implementation plan of the execution of the test.
S
2.4.10.3 Anammox batch activity test
107
70
y = -17.46 x + 72.13
2
R = 0.996
60
50
40
30
y = 4.24 x + 26.99
2
R = 0.989
20
y = -21.37 x + 69.17
2
R = 0.999
10
0
0.0
0.5
1.0
1.5
Time (h)
2.0
2.5
3.0
Figure 2.4.14 Typical output of an anammox batch activity test (type
AMX.CHE) - ammonium (●), nitrite (●) and nitrate (●) concentrations
observed during the test, resulting in depicted nitrogen conversion rates,
namely: ammonium removal rate (rAMX,NH4), nitrite removal rate (rAMX,NO2),
and nitrate production rate (rAMX,NH4_NO3). Note: data used in the figure were
obtained from another test than presented in Table 2.4.7.
The maximum volumetric ammonium and nitrite
removal, and the nitrate production rates are 17.5, 21.4
and 4.2 mg N L-1 h-1, respectively. Taking into account
the average MLVSS concentration of 2.2 g L-1 between
the sludge samples collected at the beginning and end of
108
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
ammonium and nitrite removal rate minus the maximum
volumetric nitrate production, obtaining a value of 830
mg N L-1 d-1. Similarly, the maximum specific biomass
nitrogen removal rate can be computed, which results in
a specific biomass kinetic rate (qAMX,N2 or specific
anammox activity - SAA) of 375 mg N g VSS-1 d-1. Thus,
the YNH4_NO2,AMX and YNH4_NO3,AMX ratios can be
calculated by dividing the corresponding volumetric (or
specific biomass) kinetic rates.
the test, a specific biomass ammonium oxidation rate
(qAMX,NH4_N2) of 7.9 mg N g VSS-1 h-1 can be estimated.
This rate corresponds to 189 mg N g VSS-1 d-1. Using the
same approach, the specific biomass nitrite reduction
(qAMX,NO2_N2) and nitrate production (qAMX,NH4_NO3) rates
can be calculated resulting in the maximum specific rates
of 9.7 and 1.9 mg N g VSS-1 h-1, respectively, which
corresponds to 232 and 46 mg N g VSS-1 d-1, respectively.
Finally, the maximum volumetric nitrogen removal rate
can be calculated as the sum of the maximum volumetric
Table 2.4.7 Results from the anammox batch activity test.
Anammox batch activity test
Code: AMX.CHE
Date:
Tuesday
Description:
Anammox test at 30 °C with real activated sludge
13.10.2015 10:00 h
Experimental procedure in short:
1. Confirm availability of sampling material and required equipment.
08:00
Test No.:
1
2. Confirm calibration and functionality of systems, meters and sensors.
08:10
Duration
3 h (180 min)
3. Transfer 1.0 L of sludge to batch reactor.
08:30
Substrate:
Synthetic: ammonium and nitrite (1,000 mg L each) + minerals
4. Start aerobic conditions with gentle mixing and air sparging at T and pH set points.
08:40
Sampling point:
Middle mixed liquor height in the SBR
5. 15 min before starting, take a sample for initial conditions (AMX.CHE 1).
09:40
Samples No.:
AMX.CHE 1-8
7. Start batch test: add 0.07 L of synthetic influent.
09:55
Total sample volume:
80 mL
8. Take first sample for the determination of the initial conditions (minute zero)
10:00
(5 mL normal sample, 20 mL for MLVSS)
9. Minute 30, continue sampling program according to schedule
10:30
2.5 L
10. Minute 180, stop mixing.
13:00
11. Organize the samples and clean the system.
13:10
12. Verify that all systems are swtiched off.
13:20
-1
Reactor volume:
Time (h:min)
Sampling schedule
Time (min)
Time (h)
-15
-5
0
30
60
90
120
150
180
-0.25
-0.08
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3
4
5
6
7
8
19.4
Sample No.
1
2
Parameter
ANAEROBIC PHASE
-1
8.2
-1
0
NH4-N (mg N L )
NO2 -N (mg N L )
-1
NO3 -N (mg N L )
27.5
-1
MLSS and MLVSS (mg L )
1
1
73.1
63.9
52.7
44.8
38.1
29.5
69.1
58.3
47.6
38.1
25.8
16.2
4.7
27.2
29.3
31.1
32.4
35.9
38
39.6
See table
See table
Average value of the concentration present in the synthetic substrate and and liquid phase of te sludge sample prior the start of the test
MLSS & MLVSS measurements
Sampling point
Cup No.
W1
W2
W3
W2-W1
W2-W3
MLSS
MLVSS
Ratio
1
0.09440
0.14430
0.09880
0.04990
0.04550
2,495
2,275
0.91
2
0.09480
0.14210
0.09950
0.04730
0.04260
2,365
2,130
0.90
3
0.09530
0.14190
0.09860
0.04660
0.04330
2,330
2,165
0.93
Average
2,397
2,190
0.91
2
Start test
End test
2
4
0.09350
0.14360
0.09790
0.05010
0.04570
2,505
2,285
0.91
5
0.09410
0.14330
0.09840
0.04920
0.04490
2,460
2,245
0.91
6
0.09370
0.14200
0.09850
0.04830
0.04350
2,415
2,175
0.90
Average
2,460
2,235
0.91
Sample taken before substrate addition
Biomass composition
Sampling point
-1
MLSS (mg L )
-1
MLVSS (mg L )
Ratio
-1
Ash (mg L )
Start test
End test
2,397
2,460
2,190
2,235
0.91
0.91
207
225
In this example, the observed YNH4_NO2,AMX and
YNH4_NO3,AMX ratios are 1.22 and 0.24 g N g N-1,
respectively. The obtained specific biomass nitrogen
removal rate is comparable with the kinetics reported in
Table 2.4.3, indicating a high enrichment of anammox
bacteria in the sludge sample tested. Also, the
stoichiometric parameters of interest are in good
agreement with the YNH4_NO2,AMX and YNH4_NO3,AMX
reported in Table 2.4.3, indicating that the anaerobic
ammonium oxidation is the dominant bioprocess
occurring in the sludge sample analysed. It is important
to note that when sludge samples containing COD are
ACTIVATED SLUDGE ACTIVITY TESTS
analysed, the YNH4_NO2,AMX and/or YNH4_NO3,AMX are
expected to be different from the reference stoichiometric
values due to the simultaneous occurrence of
conventional heterotrophic denitrification.
2.4.11 Additional considerations
2.4.11.1 Presence of other organisms
The presence of other microorganism rather than those
that carry out the biological conversions under
examination may alter the results of the batch tests
resulting in the under- or over-estimation of kinetics and
stoichiometric parameters of interest.
In nitrification batch activity tests, the simultaneous
occurrence of denitrifying and/or anammox activity may
lead to an incorrect estimation of nitrite removal and/or
nitrate production. Nevertheless, since both biological
processes require anoxic conditions, it is sufficient to
ensure the complete penetration of oxygen into the
biomass. While 3-4 mg O2 L-1 are considered sufficient
to ensure fully aerobic conditions when testing flocculent
and suspended sludge samples, higher oxygen
concentrations may be required when assessing the
activity of nitrifying biofilms. It is important to note that
when a batch test is performed in the presence of COD,
the aerobic COD removal performed by heterotrophic
microorganisms would further reduce the oxygen
penetration into the biofilm, thus contributing to the
creation of undesirable anoxic zones (if oxygen becomes
limiting).
In denitrification batch activity tests, the
simultaneous occurrence of anammox activity may lead
to the overestimation of nitrite and/or underestimation of
the nitrate reduction kinetics. In order to avoid this
inconvenience, the batch test should be carried out under
ammonium-limiting conditions. It is important to note
that the ammonium present during the batch activity test
should be sufficient anyway to sustain the N-source
requirements of denitrifying bacteria, which can be
calculated in advance. When the batch test aims to assess
the impact of a particular carbon source on the
denitrifying kinetics, the simultaneous occurrence of the
denitrification process carried out by microorganisms
that can store COD intracellularly (e.g. PAO and GAO)
may lead to incorrect observations. In this case, an
aeration period prior to the conduction of the batch test
can be used to completely remove the intracellular stored
COD present in the biomass.
109
In anammox batch activity tests, the simultaneous
occurrence of denitrifying and anammox activity may
lead to an incorrect estimation of the nitrite removal
and/or nitrate production kinetics, which consequently
affects the YNH4_NO2,AMX and/or YNH4_NO3,AMX. When OHO
are the dominant denitrifying population, it is sufficient
to execute the batch test in the absence of SB to limit the
presence of electron-donating compounds used for
nitrite/nitrate reduction. However, when EBPR activated
sludge is used instead, this would not be sufficient since
intracellular storage compounds would still allow the
reduction of nitrite/nitrate during the anammox batch
test. A period of aeration prior to the conduction of the
anammox batch test can be used to completely remove
the SB in the bulk and/or the intracellular stored COD
present in the activated sludge to be tested. Nevertheless,
since the anammox pathway is the only one capable of
oxidizing ammonium under anoxic conditions, anammox
activity can be satisfactorily assessed in a batch test even
in the presence of COD by following the ammonium
concentration over time.
2.4.11.2 Shortage of essential micro- and macronutrients
Though it may seem trivial, the presence of macro- and
micro-nutrients in the right concentration and (bio-)
availability is essential for the bioprocesses involved in
the nitrogen removal of activated sludge systems such as
the nitrification, denitrification and anammox processes
described in this chapter. Commonly, macro- and micronutrients are present in most municipal wastewaters, but
their presence should be checked and confirmed
particularly if the wastewater treatment plant under
examination regularly receives industrial effluents. Due
to the low yield of autotrophic microorganisms such as
nitrifiers and anammox bacteria, the lack of macronutrients is not so frequent when treating municipal
sewage. Nevertheless, it should be a point of attention
when these bioprocesses are applied to the treatment of
industrial wastewaters and in general for high-strength
wastewaters. For this purpose, a simple estimation of the
nutrient requirements of the biomass as a function of the
nitrogen load to be treated can be performed and
compared against the influent nutrient concentrations to
assess whether external addition is necessary to support
the biological growth requirements and conversion rate.
In some cases the external dosage of particular micronutrients beyond the minimal requirements can enhance
the kinetics of a specific microbial population as recently
reported in the case of iron dosage to anammox cultures
(Chen et al., 2014; Bi et al., 2014).
110
2.4.11.3 Toxicity or inhibition effects
Several compounds were identified to be toxic or
inhibitory for the bacterial communities carrying out the
nitrification, denitrification and anammox processes.
Since the abundance of OHOs capable of performing
denitrification in activated sludge systems is broad and
diverse they can acclimatise and adapt to different
environmental and operating conditions and even
withstand the presence of different potentially toxic or
inhibitory compounds. On the other hand, autotrophic
bacteria catalysing the nitrification and anammox
processes are more prone to inhibition and toxicity
issues. The list of compounds and concentration ranges
which may be toxic or inhibitory for nitrifying and
anammox bacteria is so long and complex that one should
refer to specific literature. Despite the fact that microbial
communities can also acclimatise and adapt to suboptimal conditions, the inhibitory effects of certain
compounds can lead to sub-optimal microbial process
activity and ultimately to the failure of the bioprocess. To
assess the inhibitory effect of a certain compound or of a
certain particular wastewater on the nitrifying,
denitrifying or anammox activity, a series of batch
activity tests can be conducted as described in this
chapter. The comparison between the activity measured
in the presence or absence of different concentrations of
potential inhibitory compounds can provide useful
indications about the inhibitory potential of such
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
compounds with regard to the particular activated sludge
tested. Similarly, when the wastewater used to perform
the tests is suspected to have or to generate an inhibitory
effect on a particular bioprocess, two series of batch
activity tests can be executed: one with the original
activated sludge and another one with the same biomass
but washed in a mineral solution to remove the
potentially inhibiting or toxic compounds. However, it is
important to note that such an approach may be
successful only if the inhibiting effects of the wastewater
to be tested are (rapidly) reversible.
2.4.11.4 Effects of carbon source on denitrification
It is well known that denitrification kinetics depend on
the carbon sources used as electron donors (e.g.
Mokhayeri et al., 2006, 2008). For the regular monitoring
of the denitrifying potential of activated sludge, the use
of synthetic media containing SB such as VFAs can be
good enough to provide a satisfactory assessment of
denitrification activity (as presented in this chapter). The
use and application of more complex COD sources,
which may be undoubtedly present in raw or settled
municipal wastewater, can lead to sub-optimal
denitrification activity. When external COD sources are
needed to enhance nitrogen removal (e.g. in the postdenitrification unit), the methods described in this
chapter can be used to assess the influence of different
COD sources on the denitrifying kinetics of an activated
sludge sample.
Figure 2.4.15 Execution of batch activity tests. Note the characteristic reddish colour of the enriched anammox culture (photo: Lotti, 2015).
ACTIVATED SLUDGE ACTIVITY TESTS
111
2.5 AEROBIC ORGANIC MATTER
REMOVAL
2.5.1 Process description
In conventional wastewater treatment systems
performing aerobic organic matter removal, OHOs
remove the organics present in wastewater to produce
more biomass using oxygen for respiration. From a
metabolic perspective, the removal process involves an
anabolic (for cell synthesis) and a catabolic process (to
generate the required energy for cell synthesis). In the
anabolic process, OHOs obtain the required carbon for
cell growth from the organic matter present in
wastewater. Meanwhile, in the catabolic process, an
oxidation-reduction reaction takes place involving the
transfer of electrons from the organic matter (which acts
as the electron donor) to oxygen (the electron acceptor),
generating the required energy for cell synthesis.
However, due to the rather variable mixture of
biodegradable
and
non-biodegradable
organic
compounds present in wastewater, the COD is commonly
used to estimate their total concentration. This is mostly
because the use of COD is preferred over other analytical
parameters (such as biochemical oxygen demand: BOD5
or total organic carbon: TOC) due to several advantages.
These include (Henze et al., 1997; Henze and Comeau,
2008): (i) the determination of the oxygen equivalence
(or capacity to donate electrons) of the organic
compounds, (ii) the more detailed and useful
determination of the organic strength by being able to
measure all the degradable and undegradable organics,
(iii) the potential for the balance of organics to be closed
on a COD basis (as a consequence of the previous two
advantages), as well as practical implications such as (iv)
a rapid analysis (i.e. a few hours as opposed to 5 days
required for BOD5). In general, the exact stoichiometry
involved in the aerobic removal of organics is not straight
forward. Nevertheless, the following equation for the
aerobic consumption of glucose (C6H12O6) (which
neglects most of the nutrients except nitrogen) can be
used to illustrate the biological aerobic removal process
(Metcalf and Eddy, 2003):
3C6 H12 O6 + 8O2 + 2NH3 → 2C5 H7 NO2 + 8CO2 + 14H2 O
Eq. 2.5.1
Where, C5H7NO2 is a simplified expression of the
new cells generated from the aerobic degradation of
organics (Hoover and Porges, 1952).
From a microbial growth perspective, about ⅔ of the
biodegradable organics (leading to the so-called aerobic
stoichiometric true yield, YOHO, of ~0.67 g COD-biomass
per COD-organics consumed) are converted into new
biomass via anabolism and the remaining ⅓ is oxidized
using oxygen via catabolic pathways to generate the
required energy for biomass growth (Marais and Ekama,
1976). In addition, for microbial growth, macronutrients
(like nitrogen and phosphorus) and micronutrients (such
as potassium, sodium, calcium, magnesium, zinc,
manganese and iron, among others) are needed for cell
synthesis (Metcalf and Eddy, 2003). The lack of any of
these elements can lead to limitations of the microbial
processes. It is assumed that the nitrogen and phosphorus
requirements of the new biomass produced are
approximately 0.10 g N g VSS-1 and 0.03 g P g VSS-1,
respectively (Ekama and Wentzel, 2008a). This means
that if a wastewater contains 100 mg BCOD L-1
(biodegradable organics) then the nitrogen and
phosphorus concentrations that need to be supplied
should not be lower than 4.7 mg NH4-N L-1 and 1.4 mg
PO4-P L-1, respectively, to meet the nutrient
requirements. This assumes a true yield of YOHO of 0.67,
an observed yield of 0.40 (accounting for a sludge age of
about 5 days) and a COD-to-VSS ratio of the biomass of
1.42 mg COD mg VSS-1. From a practical perspective
and to avoid nutrient limitations due to the different YOHO
observed (see Table 2.5.1), a COD:N:P ratio of 100:5:1
is usually suggested (Metcalf and Eddy, 2003).
It is important to underline that not all the organics
present in wastewater can be subject to degradation. At
the most basic classification, at least four different
organic fractions can be identified in a wastewater stream
with different physical characteristics and degree of
biodegradability that determine their removal potential in
a wastewater treatment system (Ekama and Wentzel,
2008a). These are (i) biodegradable soluble organics
rapidly converted by OHO (and thus known as readily
biodegradable organics, RBCOD or SB according to the
standardized notation (Corominas et al., 2010)), (ii)
biodegradable particulate organics that get mostly
enmeshed within the activated sludge flocs and are
subject to hydrolysis prior to biodegradation, and
therefore commonly
characterized as slowly
biodegradable organics (SBCOD or XCB in accordance
with Corominas et al., 2010), (iii) non-biodegradable
particulate organics that get mostly enmeshed in the
sludge flocs and accumulate in the activated sludge
system, and (iv) non-biodegradable soluble organics
which neither get enmeshed nor degraded and remain in
the soluble phase.
112
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
RBCOD can be rapidly utilized by OHOs. Although
the biological breakdown of SBCOD is slow, at the SRT
commonly applied in most activated sludge plants (e.g.
longer than 3-4 days), SBCOD are virtually completely
utilized (Ekama and Wentzel, 2008a). Thus, OHOs
utilize RBCOD and SBCOD for cell synthesis, producing
more biomass. The new OHO biomass generated by the
removal and conversion of RBCOD and SBCOD, as well
as the accumulation of the non-biodegradable particulate
organics, becomes part of the organic activated sludge
mass in the reactor usually measured as MLVSS.
Because of the flocculation capability of the activated
sludge, solids material is relatively highly settleable so it
can be efficiently removed in the secondary settling
tanks, providing a treated and clear effluent. The sludge
mass that settles out in the secondary settling tank is
returned to the biological reactor and eventually is
removed via the waste of activated sludge which is a
function of the plant’s SRT (Arden and Lockett, 1914).
However, since the non-biodegradable soluble organics
cannot be efficiently removed in an activated sludge
system, they leave the plant through the effluent,
contributing to the effluent COD concentration.
determination of the kinetic rates of RBCOD removal in
activated sludge and other suspended growth systems. It
does not cover the removal of SBCOD fractions because
its determination requires the execution of respirometry
tests described in Chapter 3. Moreover, and despite that
the fractionation of the influent wastewater organic
compounds into COD (as well as into N and P) is of
major importance for the design, operation, modelling
and evaluation of (activated sludge) wastewater
treatment plants, the current chapter does not aim to
elaborate on wastewater characterization and
fractionation protocols. For such a purpose, the reader is
referred to scientific and technical reports published
elsewhere (Henze, 1992; Kappeler and Gujer, 1992;
Wentzel et al., 1995; Hulsbeek et al., 2002; Roeleveld
and van Loosdrecht, 2002; Vanrolleghem et al., 2003;
WERF, 2003; Langergraber et al., 2004).
In addition to the aerobic removal of organics, in
activated sludge systems performing BNR, most of the
organic matter will be removed in the anaerobic and
anoxic stage that precedes the aerobic stage. For instance,
in the anaerobic stage of an activated sludge system
designed for EBPR, VFA are taken up by PAO, while in
the anoxic stage of a BNR plant, biodegradable organics
are used by denitrifying organisms for denitrification
purposes using nitrate or nitrite as the electron acceptor.
In addition, the potential occurrence of sulphate-reducing
processes by SRBs in the anaerobic stages of an activated
sludge system can also lead to the removal of organics.
Furthermore, the removal of organics can also take place
in anaerobic wastewater treatment systems (e.g. upflow
anaerobic sludge blankets: UASB plants) under fully
anaerobic conditions by strictly anaerobic organisms.
2.5.2 Experimental setup
As observed, organic matter removal can occur under
different environmental conditions and be performed by
different groups of microorganisms. The present section
focuses on the execution of batch activity tests to assess
aerobic organic matter removal as the primary removal
process by OHOs in conventional activated sludge
systems under fully aerobic conditions. The activity
assessment of other processes is presented in other
sections of the present and following chapters.
The present section aims to be used as a guide for the
execution of aerobic batch activity tests for the
It is important to note that some of the wastewater
characterization and fractionation protocols, as well as a
thorough determination of the actual biomass yields
using organics from different wastewaters, also need
respirometry tests (Chapter 3).
2.5.2.1 Reactors
To assess the aerobic organic matter removal activity by
activated sludge, batch tests need to be carried out under
aerobic conditions securing sufficient availability of DO
(maintaining DO concentrations higher than 2 mg L-1)
and good mixing conditions. As for other processes, it is
also important to maintain an adequate and desirable
temperature, precise pH control, and have additional
ports for sample collection and the addition of influent,
solutions, gases and any other liquid media or substrate
used in the test. In general, similar fermenters with
features and characteristics of those used for the
execution of EBPR batch activity tests can be used to
carry out the aerobic organic matter removal tests (see
Section 2.2.2.1). Similar recommendations to those
described in Section 2.2.2.1 regarding aeration, mixing
and pH control, location and characteristics of sampling
and dosing ports can also be applied here.
2.5.2.2 Activated sludge sample collection
For the execution of aerobic organic matter removal
batch activity tests, a fresh sample should be collected at
the end of the aerobic tank or phase in a sampling spot
where well-mixed conditions take place. Ideally, batch
ACTIVATED SLUDGE ACTIVITY TESTS
activity tests should be performed soon after collection
(within 2 to 3 h after sampling). If the batch activity tests
cannot be performed in situ on the same day when
sampling took place, an activated sludge sample can be
collected in a bucket or jerry can at the end of the aerobic
stage and properly transported and stored in a fridge or
using ice (to keep the temperature around 4 °C if
feasible). In any case, the in situ execution of the batch
activity tests is preferable for obvious reasons. The total
volume of activated sludge to be collected depends on the
number (repetition) of tests, reactor volume and total
volume of samples to be collected to assess the biomass
activity. Often, 10-20 L of activated sludge collected
from full-scale wastewater treatment plants can be
considered sufficient per batch. On the other hand,
samples collected from lab-scale reactors rarely provide
more than 1 L because lab-scale systems use small
reactors (from 0.5 to 2.4 L and in some cases up to 8-10
L, exceptionally 15 L) and the maximum volume that can
be withdrawn from lab-scale reactors is often set by the
daily withdrawal of the excess of sludge from the system
(which is directly related to the SRT and, consequently,
defined by the growth rate of the organism(s) of interest).
2.5.2.3 Activated sludge sample preparation
For batch activity tests performed in situ, in principle, the
sludge should be transferred from the parent reactor (in
the case of a lab-enriched sludge) or reaction tank (in the
case of pilot-scale or full-scale plants) to the fermenter or
reactor where the activity tests will take place. Then the
activated sludge should be aerated for at least 1 or 2 h to
remove any residual biodegradable COD present in the
system while the sample is also adjusted to the desired
pH and temperature of interest (as described in Section
2.2.3.5). Alternatively, to remove any residual COD the
activated sludge can be washed using the washing media
and washing procedure described in sections 2.2.3.3 and
2.2.3.5, respectively.
If only an aerobic organic removal batch activity test
is going to be executed (i.e. a nitrification test is not of
interest), then a nitrification inhibitor can be added to the
sludge sample immediately after the sludge has been
transferred to the fermenter (e.g. allyl-N-thiourea: ATU
to a recommended final concentration of 20 mg L-1). In
particular, this will restrain nitrification and consequently
avoid higher oxygen consumption if respirometry tests
are going to be executed in parallel (see Chapter 3).
Sludge samples stored under cold conditions can also be
washed with a mineral solution to remove any residual
organics.
113
For batch tests executed with sludge samples stored
under cold conditions (at around 4 °C), sludge samples
need to be 're-activated' because the cold storage
temperature slows down the bacterial metabolism. To reactivate the sludge, the activated sludge should be aerated
for 1-2 h at the desired pH and, particularly, the
temperature of study.
2.5.2.4 Media
When real wastewater (either raw or settled) is used for
the execution of activity tests, it can be fed in a relatively
straightforward manner to the reactor/fermenter. For
normal or regular conditions, the feeding step takes place
at the beginning of the test. If required, raw wastewater
can be filtered (using 1 or 2 mm sieves) or settled
(duration from 1 to 3 h). If the activity tests need to be
performed at different solids concentration, (i) the treated
effluent from the plant can be collected and used for
dilution (assuming that solids effluent concentrations are
relatively low, e.g. 20-30 mg TSS L-1), or (ii) the
activated sludge can be concentrated by decanting and
discharging the supernatant in several repeated steps until
reaching the MLSS of interest.
If different carbon sources and concentrations are to
be studied, the plant effluent can also be used to prepare
a semi-synthetic media (as long as it does not contain
toxic or inhibitory compounds) containing a RBCOD
concentration of interest which, for instance after a 1:1
dilution in the fermenter, can provide the target initial
COD concentration. However, depending upon the
nature and purpose of tests, the carbon concentrations
present in a synthetic wastewater can vary and be
adjusted proportionally to the duration of the test.
Usually, concentrations of about 400 mg COD L-1 can be
used with either lab- or full-scale activated sludge.
Besides the carbon source, the synthetic wastewater must
contain the required macro- and micro-elements. A
suggested synthetic wastewater recipe for an initial COD
of about 400 mg COD L-1 can contain per litre (Smolders
et al., 1994a): 862 mg NaAc·3H2O (400 mg COD), 107
mg NH4Cl (28 mg N), 40 mg NaH2PO4·2H2O (8 mg P),
90 mg MgSO4·7H2O, 14 mg CaCl2·2H2O, 36 mg KCl, 1
mg yeast extract and 0.3 mL of a trace element solution
(that includes per litre 10 gEDTA, 1.5 g FeCl3·6H2O,
0.15 gH3BO3, 0.03 g CuSO4·5H2O, 0.12 g MnCl2·4H2O,
0.06 g Na2MoO4·2H2O, 0.12 g ZnSO4·7H2O, 0.18 g KI
and 0.15 g CoCl·6H2O). If desired, the synthetic
wastewater can be concentrated to a higher desirable
COD concentration to account for potential dilution rates,
sterilized in an autoclave (for 1 h at 110 °C) and used as
114
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
For experiments performed with lab-enriched
cultures, it is best to execute the tests with the same
(synthetic) wastewater used for the cultivation with the
carbon concentrations of study, unless otherwise
required. Alternatively, and similar to full-scale samples,
the effluent from the reactor can be collected, filtered
through rough pore size filters to remove any coarse
particles, and used to prepare the required media.
commonly applied analytical protocols detailed in
Standard Methods (APHA et al., 2012). For the
determination of dissolved parameters like soluble COD,
PO4, NH4, NO2 and NO3, samples should be filtered
immediately after collection through 0.45 μm pore size
filters. Of the two most commonly applied methods for
the analytical determination of COD, the dichromate
method is recommended, as the permanganate method
does not fully oxidize all the organic compounds (Henze
and Comeau, 2008). The determination of VFA (like
acetate, propionate and other volatile fatty acids) can be
executed by GC. Glucose and other carbon compounds
(including VFA) can be determined by HPLC.
2.5.2.5 Analytical tests
2.5.2.6 Parameters of interest
Most of the analytical tests required (for the
determination of COD, total P, PO4, NH4 NO2, NO3, TSS,
VSS, etc.) can be performed following standardized and
To determine and assess the activity of OHOs, different
stoichiometric ratios and kinetic rates can be estimated as
displayed in Table 2.5.1.
a stock solution if several tests are going to be performed
in a defined period of time. However, the solution must
be discarded if any precipitation is observed or it loses
transparency.
Table 2.5.1 Expected stoichiometric and kinetic parameters of interest for activated sludge systems performing aerobic organic matter removal.
Parameter
Remark
Reference
Aerobic stoichiometric parameter YOHO (g COD-biomass g COD-substrate-1)
0.67
Theoretical ratio
Ekama and Wentzel (2008a)
0.72
Acetate as organic matter source
Dircks et al. (1999)
0.37
Methanol as organic matter source
McCarty (2007)
0.65
Formate as organic matter source
McCarty (2007)
0.40 -0.80; typically 0.60
g VSS g BOD-1 units
Metcalf and Eddy (2003)
0.30-0.60
g VSS g RBCOD-1 units
Metcalf and Eddy (2003)
0.67-0.792*
Different RBCOD organic matter sources
Guisasola (2005)
0.91*
Glucose as organic matter source
Dircks et al. (1999)
0.90*
Glucose as organic matter source
Goel et al. (1999)
Aerobic kinetic parameter qOHO,COD,Ox (g COD substrate g COD-biomass-1 d-1)
6
ASM2d model
Henze et al. (1999)
2-10, typically 5
g RBCOD g VSS-1 d-1
Metcalf and Eddy (2003)
3 - 10
Kappeler and Gujer (1992)
* These values deviate considerably from the theoretical yield of 0.67 gCOD gCOD-1 due to the occurrence of storage processes.
It is important to mention that an Arrhenius
temperature coefficient of between 1.060 and 1.123 (with
a typical value of 1.070) has been suggested for the
description of the aerobic organic matter removal rates
(Metcalf and Eddy, 2003).
As shown in Table 2.5.1, the theoretical
stoichiometric biomass yield on biodegradable organics
is 0.67 g COD g COD-1 (Metcalf and Eddy, 2003; Ekama
and Wentzel, 2008a). However, it is common to observe
higher stoichiometric ratios that can apparently reach up
to 0.90-0.91 g COD g COD-1 (Dircks et al., 1999; Goel
et al., 1999). If this is the case, one should be aware that
values higher than 0.67 g COD g COD-1 are caused by
the direct storage of biodegradable organics rather than a
higher biomass yield. Such processes usually occur when
RBCOD (SB) is the dominant organics in plants designed
with selectors. Further details can be found in literature
(Gujer et al., 1993; Henze et al., 2008). Regarding the
aerobic kinetic rates for organic matter removal, those
ACTIVATED SLUDGE ACTIVITY TESTS
reported in literature can vary widely from 3 to 10 g
COD-substrate g COD-biomass-1 d-1. The main reason for
this can be the net concentration of active biomass
present in the system (with respect to the total MLVSS
concentration). A short SRT (of less than 3 days) may
lead to a high fraction of active biomass with respect to
VSS present in the system which could be reflected in a
high organic matter removal rate. However, a (very) long
SRT (for instance, much higher than 20 days) will lead to
a higher accumulation of non-biodegradable VSS
(present in the influent) or produced by endogenous
respiration by OHOs (also measurable as MLVSS) which
adds to the MLVSS concentration of the sludge and leads
to a lower maximum specific COD removal rate,
qOHO,COD,Ox. Thus, the highest qOHO,COD,Ox (up to 10 g
COD-substrate g COD-biomass-1 d-1) can be expected in
tests executed with activated sludge samples containing
relatively higher concentrations of active OHO biomass,
typically observed in low sludge age systems.
2.5.3 Aerobic organic matter batch activity
tests: preparation
2.5.3.1 Apparatus
For the execution of aerobic organic batch activity tests
the following apparatus is needed:
1. A batch reactor or fermenter equipped with a mixing
system and adequate sampling ports (as described in
Section 2.5.2.1).
2. An oxygen supply (compressed air or pure oxygen
sources).
3. A pH electrode (if not included/incorporated in the
batch reactor setup).
4. A 2-way pH controller for HCl and NaOH addition
(alternatively a one-way control - generally for HCl
addition - or manual pH control can be applied
through the manual addition of HCl and NaOH).
5. A thermometer (recommended working temperature
range of 0 to 40 °C).
6. A temperature control system (if not included in the
batch reactor setup).
7. A DO meter with an electrode (if not
included/incorporated in the batch reactor setup).
8. An automatic 2-way controller for nitrogen and
oxygen gas supply (if not included in the batch
reactor setup and if tests must be performed at a
defined DO concentration).
9. A centrifuge with a working volume capacity of at
least 250 mL to carry out the sludge washing
procedure (if required).
10. A stopwatch.
115
Confirm that all the electrodes and meters (pH,
temperature and DO) are calibrated less than 24 h before
execution of the batch activity tests in accordance with
guidelines and recommendations from manufacturers
and/or suppliers.
2.5.3.2 Materials
1. Two graduate cylinders of 1 or 2 L (depending upon
the sludge volumes used) to hold the activated sludge
and wash the sludge if required.
2. At least 2 plastic syringes (preferably of 20 mL or at
least of 10 mL volume) for the collection and
determination of soluble compounds (after filtration).
3. At least 3 plastic syringes (preferably of 20 mL) for
the collection of solids, particulate or intracellular
compounds (without filtration).
4. 0.45 μm pore size filters. Preferably not of celluloseacetate because these may release certain traces of
cellulose or acetate into the collected water samples.
Consider using twice as many filters as the number of
samples that need to be filtered for the determination
of soluble compounds.
5. 10 or 20 mL transparent plastic cups to collect the
samples for the determination of soluble compounds
(e.g. soluble COD, ammonium, ortho-phosphate).
6. 10 or 20 mL transparent plastic cups to collect the
samples for the determination of mixed liquor
suspended solids and volatile suspended solids
(MLSS and MLVSS, respectively). Consider the
collection of these samples in triplicate due to the
variability of the analytical technique.
7. A plastic box or dry ice box filled with ice up to the
required volume to temporarily store (for up to 1-2 h
after the conclusion of the batch activity test) the
plastic cups and plastic tubes for centrifugation after
the collection of the samples.
8. Plastic gloves and safety glasses.
9. Pasteur or plastic pipettes for HCl and/or NaOH
addition (when the pH control is carried out
manually).
10. Metallic lab clips or clamps to close the tubing used
as a sampling port when samples are not collected
from the reactor/fermenter.
2.5.3.3 Working solutions
• Real wastewater
If real wastewater is used to carry out the batch
activity test, there is a need to collect the sample at
the influent of the wastewater treatment plant and
116
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
perform the batch activity test as soon as possible
after collection. If due to location and distance the
tests cannot be performed in less than 1 or 2 h after
collection, then one should keep the wastewater
sample cold until the test is executed (e.g. by placing
the bucket or jerry can in a fridge at 4 °C).
Nevertheless, prior to the execution of the test, the
temperature of wastewater needs to be adjusted to the
target temperature at which the batch activity test will
be executed (preferably reached in less than 1 h). A
water bath or a temperature-controlled room can be
used for this purpose.
• Synthetic influent media or substrate
If tests can be or are desired to be performed with
synthetic wastewater, then the synthetic influent can
contain a mixture of carbon and (macro- and micro-)
nutrients. Generally, they can all be mixed together in
the same media, as long as precipitation is not
observed. The usual composition and concentration
are:
a. Carbon source solution: this must be composed of
a RBCOD source like VFA (such as acetate or
propionate) or glucose, depending on the nature
or goal of the test and the corresponding research
questions. Initial COD concentration of around
400 mg L-1 is recommended for sludge samples
from either lab- or full-scale systems.
b. The nutrient-solution: this should contain all the
required macro- (ammonium, phosphorus,
magnesium, sulphate, calcium, potassium) and
micro-nutrients (iron, boron, copper, manganese,
molybdate, zinc, iodine, cobalt) to ensure that
cells are not limited by their absence and avoid
obtaining erroneous results and in extreme cases
the failure of the test. Thus, although their
concentrations may seem trivial, one must make
sure that all of the constituents are added to the
solution in the required amounts. The following
composition for a nutrient solution can be
recommended for influents containing up to 400
mg COD L-1 (based on Smolders et al., 1994a) per
litre: 107 mg NH4Cl (28 mg), 90 mg
MgSO4·7H2O, 40 mg NaH2PO4·2H2O (8 mg P),
14 mg CaCl2·2H2O, 36 mg KCl, 1 mg yeast
extract and 0.3 mL of a trace element solution
(that includes per litre 10 g EDTA, 1.5 g
FeCl3·6H2O, 0.15 gH3BO3, 0.03 g CuSO4·5H2O,
0.12 g MnCl2·4H2O, 0.06 g Na2MoO4·2H2O, 0.12
g ZnSO4·7H2O, 0.18 g KI and 0.15 g CoCl·6H2O).
Similar nutrient solutions can be used as long as
they contain all the previously reported required
nutrients.
c. Washing media: If the sludge sample must be
washed to remove any residual COD or the
presence of undesirable compounds (which may
be even inhibitory or toxic), a nutrient solution
should be prepared. The following nutrient
solution to wash the sludge could be used per litre
(Smolders et al., 1994a): 107 mg NH4Cl, 40 mg
NaH2PO4·2H2O, 90 mg MgSO4·7H2O, 14 mg
CaCl2·2H2O, 36 mg KCl, 1 mg yeast extract and
0.3 mL of a trace element solution (that includes
per litre 10 g EDTA, 1.5 g FeCl3·6H2O, 0.15
gH3BO3, 0.03 g CuSO4·5H2O, 0.12 g
MnCl2·4H2O, 0.06 g Na2MoO4·2H2O, 0.12 g
ZnSO4·7H2O, 0.18 g KI and 0.15 g CoCl·6H2O).
The washing process can be repeated two or three
times. Afterwards, the following preparation
steps of the batch activity tests can be performed.
d. ATU solution: To inhibit nitrification, an ATU
solution can be prepared to reach an initial
concentration of around 20 mg L-1 (after
addition). The ATU solution must be added
before the sludge is exposed to any aerobic
conditions (including the sludge sample
preparation or acclimatization).
e. Acid and base solutions: 100-250 mL of 0.2 M
HCl and 100-250 mL 0.2 M NaOH solutions for
automatic or manual pH control.
Finally, the required working and stock solutions to
carry out the determination of the analytical parameters
of interest must be also prepared in accordance with
Standard Methods and the corresponding protocols.
2.5.3.4 Material preparation
1. Collect the materials to execute the batch activity
tests and the definition of the number of samples to
be collected can be prepared following the same steps
presented in Section 2.2.3.4.
2. Determine the frequency of sample collection: for the
determination of the maximum RBCOD kinetic
removal rate, samples need to be collected every 5
min during the first 30-40 min of duration of the batch
activity test.
3. Collect a series of samples before any media is added
to increase the data reliability and to establish the
initial conditions of the sludge.
4. Carefully define the maximum and minimum
working volumes of the reactor according to the
suggestions presented in Section 2.2.3.4.
ACTIVATED SLUDGE ACTIVITY TESTS
5. Label all the plastic cups once the number and
frequency of collecting the samples has been defined.
Define a nomenclature and/or abbreviation to easily
identify and recognize the batch test, the sampling
time and the parameter(s) of interest to be determined
with that sample. Labelling both the plastic cup and
the cover will help to easily identify the sample.
6. Table 2.5.2 contains an example of a working sheet
to execute and keep track of the sampling collection
and batch test execution. Furthermore, it can be used
to keep a database of the different batch tests carried
out.
7. Organize all the required material within a relatively
close radius of the action around the batch setup so
any delay in handling and preparing the samples can
be avoided.
8. Calibrate all the meters (pH, DO and thermometer)
less than 24 h prior to the execution of the tests and
store them in proper solutions until the execution of
the
tests,
according
to
the
particular
recommendations of the corresponding manufacturer
or supplier. Also confirm that the readings are
reliable.
9. Store the samples properly and preserve them until
they are analysed. Table 2.2.2 provides
recommendations for sample preservation depending
on the analytical determination of the parameter of
interest.
2.5.3.5 Activated sludge preparation
These procedures consider that batch activity tests can be
performed as soon as possible after collection of samples
from full- or lab-scale systems or, in the worst case
scenario, within 24 h after collection. Ideally, the
execution of batch tests 24 h after the collection is not
recommended due to potential changes and decay of the
microorganisms. Bearing this in mind, the following
procedure is recommended to prepare the activated
sludge samples for the execution of batch activity tests:
1. If batch activity tests can be executed in less than 1 h
after collection of the sludge sample and if the sludge
sample does not need to be washed:
a. Adjust the temperature of the batch reactor to the
target temperature of the study.
b. Collect the sludge at the end of the aerobic stage.
c. Transfer the sludge sample to the reactor or
fermenter.
d. Add the ATU solution to inhibit nitrification
(particularly if respirometry tests will be
executed) to a final concentration of 20 mg L-1
(see Section 2.5.3.3).
117
e. Start mixing (100-300 rpm) and follow the
temperature and pH of the sludge sample by
placing an external thermometer inside the
reactor (if the setup does not have a built-in
thermometer) until the sludge reaches the target
temperature and pH of the study.
f. Start to aerate the sludge samples keeping a DO
concentration higher than 2 mg L-1.
g. Start the aerobic batch activity tests once the pH
and DO are stable.
2. If the activated sludge sample needs to be washed
then the steps presented in Section 2.2.3.5 (for tests
that can be executed in less than 1 h or 24 h after
collection) can be followed for the activated sludge
sample preparation using the washing media
indicated in Section 2.5.3.3.
2.5.4 Aerobic organic matter batch activity
tests: execution
Similar to previous tests, to facilitate the execution and
for data track record and archiving purposes, an
experimental implementation plan should be prepared in
advance similar to that presented in Table 2.5.2. Aerobic
organic matter removal activity tests can be executed to
assess the removal of RBCOD in activated sludge
systems. They can be performed with activated sludge
samples from full- or lab-scale systems, using real or
synthetic wastewater. Samples should be collected at the
end of the aerobic stage as long as RBCOD is not present
in the sample. The following steps are proposed for the
execution of the aerobic organic matter removal test:
Test OHO.AER.1. Single aerobic organic matter removal test
a. After the sludge has been collected, prepared and
transferred to the batch reactor (see Section 2.5.3.5),
keep the sample aerated for at least 30 min while
confirming that the pH and temperature are at the
target values of interest. Otherwise, set up the
corresponding set points (if automatic pH and
temperature controllers are applied) or adjust
manually. Wait until stable conditions are reached.
b. Ensure that the DO concentration is higher than 2 mg
L-1.
c. Keep track of the execution and sampling time with a
stopwatch.
d. Take the first samples of the water phase and biomass
once stable operating conditions are reached (around
20 min before the start of the test) to determine the
initial concentrations of the parameters of interest:
118
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
soluble COD, NH4, PO4 and MLSS and MLVSS
concentrations.
Connect the syringe to take the samples, next open or
release the lab clip or clamp that closes the sampling
port, and then pull and push the syringe several times
until a homogenous sample is collected (usually
around 5 times are required). When the syringe is full,
close the clip and remove the syringe.
Immediately filter the samples used for the
determination of soluble COD, NH4, PO4 (through
0.45 μm pore size filters). Other samples (e.g.
MLVSS and MLVSS) need to be prepared in
accordance with the corresponding protocols.
Store the samples at 4 °C in the fridge or preferably
in a cool box with ice before and during the test
execution.
Start the execution of the aerobic test at ‘time zero’
with the addition of the real or synthetic wastewater
(as the carbon source solution).
Add the real wastewater or synthetic influent to reach
an initial RBCOD-to-MLVSS ratio in the reactor of
around 0.10 mg COD mg VSS-1. For example, the
initial RBCOD and MLVSS concentrations after
mixing can range around 400 mg COD L-1 for
samples containing 4,000 mg VSS L-1. Higher or
lower ratios may also be acceptable as long as there
is sufficient time for sampling.
Ensure that the DO readings remain above 2 mg L-1
after the addition of the wastewater, and that
considerable temperature and pH variations (higher
than 1 °C or ± 0.1 for temperature and pH,
respectively) do not take place.
Duration and sampling:
(i) Usually tests can last between 2 and 4 h for initial
soluble COD concentrations of up to 400 mg
COD L-1 (at pH 7.0 and 20 °C).
(ii) To determine the aerobic kinetic parameters,
samples for the determination of soluble COD
should be collected every 5 min in the first 30-40
min of execution of the test. After this period, the
sampling frequency can be reduced to 10 or 15
min during the first 1 h, and later on to every 15
or 30 min until the test is finished.
Conclude the aerobic test with the collection of
samples for the determination of soluble COD,
MLSS, MLVSS, NH4 and PO4 concentrations.
Organize the samples and ensure that all the samples
are complete and properly labelled to avoid mixing
the samples and trivial mistakes.
Preserve and store the samples as recommended by
the corresponding analytical procedures until the
collected samples are analysed.
o. Clean up the apparatus and take appropriate measures
to maintain and look after the different sensors,
equipment and materials.
p. Keep (part of) the sludge used in the test for possible
further use (e.g. for microbial identification, see
chapters 7 and 8).
2.5.5 Data analysis
The growth yield of biomass on COD is the main
stoichiometric parameter of interest to assess the aerobic
stoichiometry of OHO (Table 2.5.1). However, the
aerobic batch activity test presented in this section cannot
be used for the determination of YOHO. This is mostly
because, for the determination of YOHO, respirometry tests
need to be executed in parallel to the aerobic tests to
assess the amount of COD being used for energy
generation. This is presented elsewhere (Wentzel et al.,
1995; Dircks et al., 1999; Goel et al., 1999; Guisasola et
al., 2005).
The maximum specific aerobic RBCOD removal rate
(qOHO,COD,Ox) can be computed by plotting the
experimental data (Y-axis) versus time (X-axis) and
fitting the experimental data obtained in the aerobic batch
activity tests using linear regression. Because one is
interested in the maximum rates, a linear regression
approach can be applied by fitting more than 4 to 5
experimental data points while achieving a statistical
determination coefficient (R2) not lower than 0.90-0.95.
This is the main reason why the sampling frequency
within the first 30-40 min of execution of the batch
activity tests is set to 5 min. The maximum volumetric
kinetic rate can be determined based on the slope of the
linear regression equation. This will result in the
determination of the maximum volumetric rate (usually
reported in units such as mg L-1 h-1 or g m-3 d-1). Figure
2.5.1 illustrates an estimation of the maximum aerobic
kinetic removal rates for an OHO culture. The rate can be
expressed as the maximum specific kinetic rate by
dividing the volumetric rate (or value of the slope) by the
concentration of activated sludge VSS. It is important to
note that the maximum biomass-specific growth rate
cannot be computed in a straightforward manner by
following the increase in biomass concentrations during
the cycle since it may be practically negligible and fall
into the standard error of the analytical determination of
MLVSS. Instead, either long-term continuous tests can
be executed or the aerobic activity tests need to be
combined with respirometry tests (Chapter 3) as
presented elsewhere in literature (Kappeler and Gujer,
1992; Wentzel et al., 1995).
COD (mg L-1)
ACTIVATED SLUDGE ACTIVITY TESTS
rOHO,COD,Ox (mg COD L-1 h-1)
119
every 5 min in the first 30 min of the aerobic phase.
Immediately after collection, all the samples were
prepared and preserved as described in Section 2.5.3.4.
All the collected samples were analysed as described in
Section 2.5.2.4. Table 2.5.2 shows the experimental
implementation plan of the test.
2.5.6.2 Data analysis
Following the results of the batch activity test shown in
Table 2.5.2, Figure 2.5.2 displays the estimation of the
maximum volumetric OHO kinetic rates by applying
linear regression.
Time (h)
Figure 2.5.1 Example of the determination of the maximum aerobic
volumetric kinetic rate for organic matter removal (expressed in mg COD L-1
h-1) in an aerobic batch activity test.
250
2.5.6 Example
2.5.6.1 Description
To illustrate the execution of an aerobic batch activity
test for OHO, data from a test performed at 15 °C with a
full-scale activated sludge sample is presented in this
section. The Test OHO.AER.1 was carried out to
determine the aerobic kinetic rate of OHO for RBCOD
removal. Thus, the batch activity test was performed
using a 3.0 L reactor. All the equipment, apparatus and
materials were prepared as described in Section 2.5.3.
The pH and DO sensors were calibrated less than 24 h
before the test execution. The test lasted 4 h. Prior to the
batch test, 1.25 L of fresh activated sludge collected in
the end of the aerobic phase of a full-scale plant was
transferred to the reactor and acclimatized for 1 h at 15
°C under slow mixing (100 rpm) at pH 7.0 following the
recommendations described in Section 2.5.3.5. Activated
sludge preparation for tests was performed less than 1 h
after sludge collection. Thereafter, 20 min before the start
of the test, mixing was increased to 300 rpm and samples
for the determination of the parameters of interest were
collected (in accordance with the execution of Test
OHO.AER.1). The test started with the addition of 1.25
L synthetic media containing 400 mg COD L-1 as Ac.
Other macro- and micro-nutrients as well as 20 mg L-1
ATU were included in synthetic media in accordance
with Section 2.5.3.3. Because the test was executed at 15
°C, the temperature of the synthetic media was also
adjusted to 15 °C in a water bath operated at the same
temperature before addition. Samples were collected
COD (mg L-1)
200
Y = 378.86 X + 234.33
R 2= 0.998
150
100
50
0
0.0
0.5
1.0
1.5
2.0 2.5
Time (h)
3.0
3.5
4.0
4.5
Figure 2.5.2 Graphic representation of the data obtained in the example
of an experimental implementation plan for the execution of a batch
activity test (Type OHO.AER.1) performed with a full-scale activated sludge
at 15 °C using synthetic influent at pH 7.0. The main trend line shows the
RBCOD conversion rate for the further estimation of the maximum specific
kinetic rate.
Based on the maximum volumetric COD removal
rate showed in Figure 2.5.2 and taking into account the
MLVSS concentration of 1,890 mg L-1 (as an average
MLVSS concentration observed between sludge samples
collected at the beginning and end of the test), a specific
RBCOD removal rate of 0.20 g COD g VSS-1 h-1 can be
estimated, which corresponds to about 4.81 g RBCOD g
VSS-1 d-1. Using a COD-to-VSS ratio for the biomass of
1.42 g COD g VSS-1 (Metcalf and Eddy, 2003), a
qOHO,COD,Ox of 3.39 g RBCOD g VSS-1 d-1 can be
computed. Furthermore, since the test was not executed
at 20 °C but at 15 °C, using a typical Arrhenius
coefficient of 1.07 for OHO activity (Metcalf and Eddy,
2003), a qOHO,COD,Ox of 4.75 g RBCOD g COD-1 d-1 can be
determined (3.39 g RBCOD g VSS-1 d-1 / 1.07(15-20)). This
120
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
noticed that: (i) the initial COD concentration in the
activated sludge prior to the wastewater addition was
around 56 mg L-1 and, (ii) the final COD concentration at
the end of the test remained at around 30 mg COD L-1.
The latter concentration corresponds to the soluble nonbiodegradable COD present in the activated sludge
sample which cannot be removed by biological means (as
explained earlier in this chapter).
value is in the range of other maximum specific kinetic
rates reported for aerobic organic matter removal
processes. It is important to note that the synthetic
influent contained 400 mg COD L-1, and the COD
concentration measured 5 min after the start of the test
was 203 mg COD L-1 due to the dilution of the 1.25 L of
synthetic influent with the 1.25 L of activated sludge
(which did not contain RBCOD) and the rapid kinetic
removal rate of the biomass. Moreover, it can also be
Table 2.5.2 Example of an experimental implementation plan for the execution of a batch activity test (Type OHO.AER.1) performed with a full-scale activated
sludge at 15 °C using synthetic influent at pH 7.0.
Aerobic COD removal batch tests
Code: OHO.AER.1
Date:
Description:
Test No.:
Duration
Thursday
09.10.2015 10:00 h
Tests at 25 °C, pH 7, synthetic substrate and full-scale sample
1 of 6
4,0 h (240 min)
Substrate:
Sampling point:
Samples No.:
Total sample volume:
Synthetic: Acetate (350 mg L ) + mineral solution with N and P
Middle mixed liquor height in the SBR
OHO.AER.1(1-22)
222 mL
(10 mL for MLVSS, 6 mL for other samples)
2.5 L
-1
Reactor volume:
Sampling schedule
Time (min)
Time (h)
Sample No.
Parameter
-20
-0.33
1
-1
0
0.00
1
HAc (mg L )
56
1
1.2
-1
PO 4 -P (mg L )
-1
5.3
NH4 -N (mg L )
-1
MLSS and MLVSS (mg L )
1
Experimental procedure in short:
1. Confirm the availability of sampling material and required equipment.
2. Confirm calibration and functionality of systems, meters and sensors.
3. Transfer 1.25 L of sludge to batch reactor.
4. Start with gentle mixing and air sparging at T and pH set points.
5. 20 min before starting take sample for initial conditions (OHO.AER.1.1).
7. Start the test: add 1.25 L synthetic influent (minute zero).
8. Minute 5, continue sampling program according to schedule.
9. Minute 240, stop aeration and mixing.
10. Organize the samples and clean the system.
11. Verify that all systems are swtiched off.
5
0.08
2
10
0.17
3
15
0.25
4
203
195.5
184
20
25
0.33
0.42
5
6
AEROBIC PHASE
173
158.5
08:40
09:40
10:00
10:05
14:00
14:15
14:20
30
0.50
7
40
0.67
8
50
0.83
9
60
1.00
10
75
1.25
11
143.5
135.5
129
119.5
83
1
See table
See table
Average value of the concentration present in the synthetic substrate and and liquid phase of the sludge sample prior to the start of the test.
Sampling schedule (continued)
Time (min)
Time (h)
Sample No.
Parameter
-1
HAc (mg L )
90
1.50
12
105
1.75
13
120
2.00
14
135
2.25
15
73
63
43
38
150
165
2.50
2.75
16
17
AEROBIC PHASE
28
30.5
180
3.00
18
195
3.25
19
210
3.50
20
225
3.75
21
33
25.5
28
26.5
-1
PO 4 -P (mg L )
-1
MLSS & MLVSS measurements
Sampling point
See table
Cup No.
W1
W2
W3
W2-W1
W2-W3
MLSS
MLVSS
Ratio
Start aerobic phase
1
2
3
0.08835
0.08835
0.08834
0.10741
0.10759
0.10683
0.08849
0.09018
0.08940
0.01906
0.01924
0.01849
End aerobic phase
4
5
6
0.08868
0.08764
0.08722
0.10758
0.10617
0.10648
0.08934
0.08874
0.08973
0.01890
0.01853
0.01926
0.01892
0.01742
0.01742
Average
0.01824
0.01742
0.01675
Average
1,906
1,924
1,849
1,893
1,890
1,853
1,926
1,890
1,892
1,742
1,742
1,792
1,824
1,742
1,675
1,747
0.99
0.91
0.94
0.95
0.97
0.94
0.87
0.93
2
Sample taken before substrate addition.
Biomass composition
Sampling point
-1
Ratio
-1
Ash (mg L )
30
8.3
-1
MLSS and MLVSS (mg L )
MLSS (mg L )
-1
MLVSS (mg L )
240
4.00
22
1.8
NH4 -N (mg L )
2
Time (h:min)
08:00
08:10
08:30
Start Aer.
End Aer.
1,893
1,890
1,792
1,747
0.95
0.93
Orthophosphate (PO4 -P)
101
142
Ammonium (NH4 -N)
Note:
Acetate (CH2 O)
3-
+
30.03 mg C-mmol
-1
31.00 mg P-mmol
-1
14.00 mg N-mmol
-1
ACTIVATED SLUDGE ACTIVITY TESTS
121
The presence of nitrogen (as ammonia) and
phosphorus (as orthophosphate) at the beginning (of
around 5.3 and 1.2 mg L-1 in the activated sludge,
respectively) and end of the test (of 8.3 and 1.8 mg L-1,
accordingly) indicates that these micronutrients were not
limiting. Actually, the concentrations are higher at the
end of the test because the synthetic influent contained
around 28 mg NH4-N L-1 and 8 mg PO4-P L-1. Thus, after
dilution and nutrient consumption for biomass synthesis,
8.3 mg NH4-N L-1 and 1.8 mg PO4-P L-1 remained at the
end of the aerobic test. Should the actual nitrogen (Nreq)
and phosphorus (Preq) requirements be known, then they
can be estimated with the following expressions:
Nreq =
YCOD · RBCODremoved · NS
fCV
Eq. 2.5.2
Preq =
YCOD · RBCODremoved · PS
fCV
Eq. 2.5.3
Where, YOHO is the biomass growth yield (in CODbiomass COD-substrate-1), RBCODremoved is the
concentration of RBCOD removed in the test (taking into
account the initial COD concentration and the potential
dilution effects), NS is the nitrogen requirement for
biomass growth assumed to be around 0.10 g N g VSS-1,
PS is the phosphorus requirement for biomass growth
assumed to be around 0.03 g P g VSS-1, fCV is the CODto-VSS ratio of the sludge, commonly assumed to be 1.42
or 1.48 g COD g VSS-1 (Metcalf and Eddy, 2003; Ekama
and Wentzel, 2008a).
2.5.7 Additional considerations and
recommendations
2.5.7.1 Simultaneous storage and microbial growth
Different studies performed in lab- and full-scale systems
have documented the simultaneous occurrence of
RBCOD storage and microbial growth under aerobic
conditions (van Loosdrecht et al., 1997; Beun et al.,
2000; Dircks et al., 2001; Martins et al., 2003; Sin et al.,
2005). Under such conditions and when OHOs are
exposed to high substrate gradients (like those observed
in plug-flow reactors or in aerobic selectors in activated
sludge systems), part of the RBCOD present in the bulk
liquid is transported through the cell membrane and
stored as intracellular polymers like PHA for their further
use for microbial growth when the substrate
concentrations are low or exhausted (van Loosdrecht et
al., 1997; Beun et al., 2000, Dircks et al., 2001); the
remaining RBCOD fraction that is not intracellularly
stored is directly used for biomass growth (Sin et al.,
2005). Such a combination of processes will lead to
apparently higher YOHO values since the aerobic RBCOD
storage processes apparently require less oxygen than the
direct growth on RBCOD (at least when the substrate
storage process takes place). The role and importance of
storage processes in activated sludge systems has been
significantly recognized to such a degree that different
models have been developed and discussed to provide a
better description of the simultaneous occurrence of
storage and growth in full-scale systems (Gujer et al.,
1999; Sin et al., 2005; Guisasola et al., 2005, van
Loosdrecht et al., 2015). For the execution of aerobic
matter removal tests focused on the determination of the
kinetic removal rates, the simultaneous occurrence of
these processes should not have direct practical
implications. However, the reader should be aware that
they will influence the oxygen uptake rate profiles and
lead to certain deviations in the suggested YOHO values as
thoroughly described elsewhere (Gujer et al., 1999).
2.5.7.2 Lack of nutrients
Though it may seem trivial, the presence of macro- and
micro-nutrients in the right concentrations is essential for
successful operation of an activated sludge system. These
are usually present in sufficient amounts in most
municipal wastewaters, but their presence should be
checked and confirmed particularly if the sewage plant is
regularly receiving industrial effluents. For this purpose,
a simple estimation of the nutrient requirements of the
biomass as a function of the COD removal concentration
(as suggested in this chapter) can be performed and
compared with the influent nutrient concentrations to
assess whether there are sufficient nutrients available to
cover the biological growth requirements. Lack of such
nutrients can lead to severe problems including pin floc
and filamentous bulking sludge (Eikelboom, 2000;
Martins, 2004), affecting the effluent quality and, in
extreme cases, can lead to the failure of the activated
sludge system.
2.5.7.3 Toxicity or inhibition
The abundance of OHOs in activated sludge systems is
so broad and diverse that they can acclimatise and adapt
to different environmental and operating conditions and
even stand the presence of different potentially toxic or
inhibitory compounds, particularly proceeding from
industrial activities. Although they can acclimatise and
adapt to such diverse conditions, the inhibitory effects
can lead to sub-optimal microbial process activity. To
122
assess such potential effects, two batch activity tests can
be executed: one with the original sludge and another one
with the same sludge but washed in a mineral solution to
remove potentially inhibiting or toxic compounds (as
explained in Section 2.5.3.5). To compare the results
obtained from each batch activity test, it will be very
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
useful to assess whether the presence of certain
compounds inhibits the activity of OHOs. However, one
should be aware that such an approach may only be
successful if the inhibiting effects are (rapidly)
reversible.
Figure 2.5.3 A high-tech bioreactor system for batch and continuous tests with microbial cultures (photo: Applikon Biotechnology B.V., 2016)
ACTIVATED SLUDGE ACTIVITY TESTS
123
Annex I: Unit conversion coefficients
Table A1 Conversion factors (CF) for net-conversions, stoichiometric and kinetic parameters for carbon, phosphorus, nitrogen and sulphur.
Parameter
ANAEROBIC PARAMETERS
Phosphate released/HAc uptake ratio
Glycogen utilization/HAc uptake ratio
PHB formation/HAc uptake ratio
PHV formation/HAc uptake ratio
PH2MV formation/HAc uptake ratio
Phosphate release/HPr uptake ratio
Glycogen utilization/HPr uptake ratio
PHB formation/HPr uptake ratio
PHV formation/HPr uptake ratio
PH2MV formation/HPr uptake ratio
PHV formation/PHB formation
Sulphate reduction/HAc uptake ratio
Sulphate reduction/HPr uptake ratio
Active biomass formation/Sulphate reduction
ANOXIC PARAMETERS
PHB degradation/Nitrate removal
PHV degradation/Nitrate removal
PH2MV degradation/Nitrate removal
Glycogen formation/Nitrate removal
Poly-P formation/Nitrate removal
Active biomass formation/Nitrate removal
Methanol degradation/Nitrate removal
Ethanol degradation/Nitrate removal
Acetate degradation/Nitrate removal
PHB degradation/Nitrite removal
PHV degradation/Nitrite removal
Anoxic PH2MV degradation/Nitrite removal
Glycogen formation/Nitrite removal
Poly-P formation/Nitrite removal
Active biomass formation/Nitrite removal
Methanol degradation/Nitrite removal
Ethanol degradation/Nitrite removal
Acetate degradation/Nitrite removal
Nitrite removal/Ammonium removal
Nitrate removal/Ammonium removal
Act. biom. formation/Ammonium cons.
AEROBIC PARAMETERS
PHB degradation/Oxygen consumption
PHV degradation/Oxygen consumption
PH2MV degradation/Oxygen consumption
Glycogen formation/Oxygen consumption
Poly-P formation/Oxygen consumption
Active biom. formation/Oxygen cons.
Active biomass formation/Ammonium cons.
Unit (mg basis)
Unit (mole basis)
CF
Unit (COD basis)
CF
mg PO4 mg C2H4O2-1
mg (C6H10O5)n mg C2H4O2-1
mg (C4H6O2)n mg C2H4O2-1
mg (C5H8O2)n mg C2H4O2-1
mg (C6H10O2)n mg C3H6 O2-1
mg PO4 mg C3H6 O2-1
mg (C6H10O5)n mg C3H6 O2-1
mg (C4H6O2)n mg C3H6 O2-1
mg (C5H8O2)n mg C3H6 O2-1
mg (C6H10O2)n mg C3H6 O2-1
mg (C5H8O2)n mg (C4H6O2)n-1
mg SO4 mg C2H4O2-1
mg SO4 mg C3H6 O2-1
mg CH2.09O0.54N0.20P0.02 mg SO4-1
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
P-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
C-mol C-mol-1
S-mol C-mol-1
S-mol C-mol-1
C-mol S-mol-1
0.32
1.11
1.40
1.50
1.58
0.80
0.92
1.15
1.23
1.30
1.08
0.31
0.26
3.70
mg PO4 mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg PO4 mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg COD mg COD-1
mg SO4 mg COD-1
mg SO4 mg COD-1
mg COD mg SO4-1
0.95
1.12
1.57
1.82
2.01
0.65
0.77
1.08
1.25
1.39
1.16.
0.95
0.65
1.52
mg (C4H6O2)n mg NO3-1
mg (C5H8O2)n mg NO3-1
mg (C6H10O2)n mg NO3-1
mg (C6H10O5)n mg NO3-1
mg (PO3Mg0.33K0.33)n mg NO3-1
mg (CH2.09O0.54N0.20P0.02) mg NO3
mg CH4O mg NO3-1
mg C2H6O mg NO3-1
mg C2H4O2 mg NO3-1
mg (C4H6O2)n mg NO2-1
mg (C5H8O2)n mg NO2-1
mg (C6H10O2)n mg NO2-1
mg (C6H10O5)n mg NO2-1
mg (PO3Mg0.33K0.33)n mg NO2-1
mg CH2.09O0.54N0.20P0.02 mg NO2-1
mg CH4O mg NO2-1
mg C2H6O mg NO2-1
mg C2H4O2 mg NO2-1
mg NO2 mg NH4-1
mg NO3/mg NH4-1
mg(CH2.09O0.54N0.20P0.02) mg NH4-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
P-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
P-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
C-mol N-mol-1
N-mol N-mol-1
N-mol N-mol-1
C-mol N-mol-1
2.88
3.10
3.26
2.30
0.62
2.38
1.92
2.70
2.08
2.14
2.30
2.42
1.71
0.46
1.77
1.43
2.00
1.54
0.39
0.29
0.69
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO3-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg COD mg NO2-1
mg NO2 mg NH4-1
mg NO3 mg NH4-1
mg COD mg NH4-1
1.66
1.92
2.12
1.18
1.52
1.98
2.06
1.06
1.66
1.92
2.12
1.18
1.52
1.98
2.06
1.06
-
mg (C4H6O2)n mg O2-1
mg (C5H8O2)n mg O2-1
mg (C6H10O2)n mg O2-1
mg (C6H10O5)n mg O2-1
mg (PO3Mg0.33K0.33)n mg O2-1
mg (CH2.09O0.54N0.20P0.02) mg O2-1
mg (CH2.09O0.54N0.20P0.02) mg NH4-1
C-mol mol O2-1
C-mol mol O2-1
C-mol mol O2-1
C-mol mol O2-1
P-mol mol O2-1
C-mol mol O2-1
C-mol N-mol-1
0.37
0.40
0.42
0.29
0.08
0.31
0.69
mg COD mg O2-1
mg COD mg O2-1
mg COD mg O2-1
mg COD mg O2-1
mg (PO3Mg0.33K0.33)n mg O2-1
mg COD mg O2-1
mg COD mg NH4-1
1.66
1.92
2.12
1.18
1.52
1.52
124
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
GENERAL PARAMETERS
NA
mg P mg C-1
P-mol C-mol-1
-1
NA
mg C mg C
C-mol C-mol-1
-1
NA
mg C mg N
C-mol N-mol-1
-1
NA
mg P mg N
P-mol N-mol-1
-1
NA
mg C mg O2
C-mol O2-mol-1
-1
NA
mg P mg O2
P-mol O2-mol-1
MOLECULAR WEIGHTS
HF
mg CH2O2
C-mol
HAc
mg C2H4O2
C-mol
HPr
mg C3H6 O2
C-mol
HBr
mg C4H8O2
C-mol
Lactate
mg C3H6O3
C-mol
Methanol
mg CH4O
C-mol
Ethanol
mg C2H6O
C-mol
Glucose
mg C6H12O6
C-mol
Sulphide
mg S2S-mol
Carbondioxide
mg CO2
C-mol
Phosphate
mg PO4
P-mol
Nitrogen
mg N2
N-mol
Ammonium
mg NH4
N-mol
Nitrate
mg NO3
N-mol
Nitrite
mg NO2
N-mol
Oxygen
mg O2
mol O2
Sulphate
mg SO4
S-mol
PHB
mg (C4H6O2)n
C-mol
PHV
mg (C5H8O2)n
C-mol
PH2MV
mg (C6H10O2)n
C-mol
Glycogen
mg (C6H10O5)n
C-mol
Poly-P
mg (PO3Mg0.33K0.33)n
P-mol
Active biomass
mg (CH2.09O0.54N0.20P0.02)
C-mol
ADDITIONAL FACTORS
Common active biomass fraction of VSS in
mg VSS
mg active biomass
EBPR enrichment cultures a
a
Glycogen content of the sludge varies, dependent on the poly-P content (Welles et al., 2015b).
0.39
1.00
1.16
0.45
1.33
1.04
0.022
0.033
0.041
0.045
0.033
0.031
0.043
0.033
0.031
0.023
0.011
0.071
0.055
0.048
0.065
0.031
0.010
0.046
0.050
0.053
0.037
0.010
0.038
0.80
-
-
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
mg COD
0.35
1.06
1.53
1.80
1.06
1.98
2.06
1.06
2.00
1.66
1.92
2.12
1.18
1.52
-
References
Abdeen, S., Di, W., Hui, L., Chen, G.-H. and van Loosdrecht,
M.C.M. (2010). Fecal coliform removal in a sulfate reducing
autotrophic denitrification and nitrification integrated (SANI)
process for saline sewage treatment. Water Science and
Technology, 62(11): 2564-2570.
Acevedo, B., Oehmen, A., Carvalho, G., Seco, A., Borrás, L. and
Barat, R. (2012). Metabolic shift of polyphosphateaccumulating organisms with different levels of polyphosphate
storage. Water Research, 46(6): 1889-1900.
Acevedo, B., Borrás, L., Oehmen, A. and Barat, R. (2014).
Modelling the metabolic shift of polyphosphate-accumulating
organisms. Water Research, 65: 235-244.
Al Abbas, F.M., Williamson, C., Bhola, S.M., Spear, J.R., Olson,
D.L., Mishra, B. and Kakpovbia, A.E. (2013). Influence of
sulfate reducing bacterial biofilm on corrosion behavior of lowalloy, high-strength steel (API-5L X80). International
Biodeterioration & Biodegradation, 78: 34-42.
Ahn, J., Schroeder, S., Beer, M., McIlroy, S., Bayly, R.C., May,
J.W., Vasiliadis, G. and Seviour, R.J. (2007). Ecology of the
microbial community removing phosphate from wastewater
under continuously aerobic conditions in a sequencing batch
bioreactor. Applied and Environmental Microbiology, 73(7):
2257-2270.
Almeida, P.F., Almeida, R.C.C., Carvalho, E.B., Souza, E.R.,
Carvalho, A.S., Silva, C.H.T.P. and Taft, C.A. (2006).
Overview of sulfate-reducing bacteria and strategies to control
biosulfide generation in oil waters. In: Modern biotechnology
in medicinal chemistry and industry, Taft, C.A. (Ed.). Research
Signpost, Trivandrum, India.
American Public Health Association (APHA). American Water
Works Association (AWWA), and Water Environment
Federation (WEF). (2012). Standard methods for the
examination of water and wastewater. 22nd edition, ISBN:
0875530133, Washington, D.C.
ACTIVATED SLUDGE ACTIVITY TESTS
Arden, E. and Lockett, W.T. (1914). Experiments on the oxidation
of sewage without the aid of filters. Journal of the Society of
Chemical Industry, 33(10): 523-539.
Artiga, P., Gonzalez, F., Mosquera-Corral, A., Campos, J.L.,
Garrido, J.M., Ficara, E. and Méndez, R. (2005). Multiple
analyses reprogrammable titration analyser for the kinetic
characterisation of nitrifying and autotrophic denitrifying
biomass. Biochemical Engineering Journal, 26: 176-183.
Baetens, D., Aurola, A.M., Foglia, A., Dionisi, D. and van
Loosdrecht, M.C.M. (2002). Gas chromatographic analysis of
polyhydroxybutyrate in activated sludge: a round-robin
test. Water Science and Technology, 46(1-2): 357-361.
Bale, C.W., Chartrand, P., Degtrov, S.A., Eriksson, G., Hack, K.,
Ben Mahfoud, R., Melançon, J., Pelton, A.D. and Petersen, S.
(2002). FactSage thermochemical software and databases.
Calphad, 26: 189-228.
Balk, M., Altınbaş, M., Rijpstra, W.I.C, Sinninghe Damsté, J.S. and
Stams, A.J.M. (2008). Desulfatirhabdium butyrativorans gen.
nov., sp. nov., a butyrate-oxidizing, sulfate-reducing bacterium
isolated from an anaerobic bioreactor. International Journal of
Systematic and Evolutionary Microbiology, 58: 110-115.
Barañao, P.A. and Hall, E.R. (2004). Modelling carbon oxidation in
CTMP pulp mill activated sludge systems: calibration of
ASM3. Water Science and Technology, 50(3): 1-10.
Barat, R., Montoya, T., Borras, L., Ferrer, J. and Seco, A. (2008).
Interactions between calcium precipitation and the
polyphosphate-accumulating bacteria metabolism. Water
Research, 42(13): 3415-3424.
Barker, P.S. and Dold, P.L. (1995). COD and nitrogen mass
balances in activated-sludge systems. Water Research, 29(2):
633-643.
Bettazzi, E., Caffaz, S., Vannini, C. and Lubello, C. (2010). Nitrite
inhibition and intermediates effects on Anammox bacteria: a
batch-scale experimental study. Process Biochemistry, 45(4):
573-580.
Beun, J.J., Paletta, F., van Loosdrecht, M.C.M. and Heijnen, J.J.
(2000). Stoichiometry and kinetics of Poly-B-hydroxybutyrate
metabolism in aerobic, slow growing activated sludge cultures.
Biotechnology and Bioengineering, 67: 379-389.
Bi, Z., Qiao, S., Zhou, J., Tang, X. and Zhang, J. (2014). Fast startup of Anammox process with appropriate ferrous iron
concentration. Bioresource Technology, 170: 506-512.
Bilanovic, D., Battistoni, P., Cecchi, F., Pavan, P. and MataAlvarez, J. (1999). Denitrification under high nitrate
concentration and alternating anoxic conditions. Water
Research, 33(15): 3311-3320.
Bjerrum, J., Schwarzenbach, G. and Sillén, L.G. (1958). Stability
Constants. Chemical Society, London.
Blackburne, R., Vadivelu, V.M., Yuan, Z. and Keller, J. (2007).
Determination of growth rate and yield of nitrifying bacteria by
measuring carbon dioxide uptake rate. Water Environment
Research, 79(12): 2437-2445.
Bock, E., Schmidt, I., Stuven, R. and Zart, D. (1995). Nitrogen loss
caused by denitrifying Nitrosomonas cells using ammonium or
hydrogen as electron donors and nitrite as electron acceptor.
Archives of Microbiology, 163: 16-20.
Bogaert, H., Vanderhasselt, H., Gernaey, K., Yuan, Z., Thoeye, C.
and Verstraete, W. (1997). A new sensor based on pH-effect of
the denitrification process. Journal of Environmental
Engineering, 123: 884-891.
Boyles, S. (1997). The Science of Chemical Oxygen Demand.
Technical Information Series, Booklet No. 9, HACH Company,
USA, pp: 1-23.
Brdjanovic, D., Hooijmans, C.M., van Loosdrecht, M.C.M.,
Alaerts, G.J. and Heijnen, J.J. (1996). The dynamic effects of
potassium limitation on biological phosphorus removal. Water
Research, 30(10): 2323-2328.
125
Brdjanovic, D., van Loosdrecht, M.C.M., Hooijmans, C.M.,
Alaerst, G. J. and Heijnen, J. J. (1997). Temperature effects on
physiology of biological phosphorous removal systems. ASCE
Journal of Environmental Engineering, 123: 144-154.
Brdjanovic, D. (1998a). Modelling biological phosphorous removal
in activated sludge systems. PhD Thesis. Delft University of
Technology, ISBN: 9054104155, Balkema Publishers,
Rotterdam, the Netherlands.
Brdjanovic, D., Logemann, S., van Loosdrecht, M.C.M.,
Hooijmans, C.M., Alaerts, G.J. and Heijnen, J.J. (1998b).
Influence of temperature on biological phosphorus removal:
process and molecular ecological studies. Water Research,
32(4): 1035-1048.
Brdjanovic, D., Slamet, A., van Loosdrecht, M.C.M., Hooijmans,
C.M., Alaerts, G.J. and Heijnen, J.J. (1998c). Impact of
excessive aeration on biological phosphorus removal from
wastewater. Water Research, 32(1): 200-208.
Brdjanovic, D., van Loosdrecht, M.C.M., Veersteeg, P., Hooijmans,
C.M., Alaerts, G.J. and Heijnen, J.J. (2000). Modelling COD,
N and P removal in a full-scale WWTP Haarlem
Waarderpolder. Water Research, 34: 846-858.
Burow, L.C., Kong, Y., Nielsen, J.L., Blackall, L.L. and Nielsen, P.H.
(2007). Abundance and ecophysiology of Defluviicoccus spp.,
glycogen-accumulating organisms in full-scale wastewater
treatment processes. Microbiology, 153(1): 178-185.
Butlin, K.R., Selwyn, S.C. and Wakerley, D.S. (1956). Sulfide
production from sulphate-enriched sewage sludges. Journal of
Applied Bacteriology, 19(1): 3-15.
Caccavo, F.Jr., Frolund, B., van Ommen Kloeke, F. and Nielsen,
P.H. (1996). Deflocculation of activated sludge by the
dissimilatory Fe(III)-reducing bacterium Shewanella alga BrY.
Applied and Environmental Microbiology, 62: 1487-1490.
Carvajal-Arroyo, J.M., Sun, W., Sierra-Alvarez, R. and Field, J.A.
(2013). Inhibition of anaerobic ammonium oxidizing
(Anammox) enrichment cultures by substrates, metabolites and
common wastewater constituents. Chemosphere, 91: 22-27.
Carvalho, G., Lemos, P.C., Oehmen, A. and Reis, M.A. (2007).
Denitrifying phosphorus removal: linking the process
performance with the microbial community structure. Water
Research, 41(19): 4383-4396.
Carvalheira, M., Oehmen, A., Carvalho, G., Eusébio, M. and Reis,
M.A. (2014). The impact of aeration on the competition
between polyphosphate accumulating organisms and glycogen
accumulating organisms. Water Research, 66: 296-307.
Cao, Y., Ang, C., Chua, K., Woo, F., Chi, H., Bhawna, B., Chong,
C.T., Ganesan, N., Ooi, K.E. and Wah, Y. (2009). Enhanced
biological phosphorus removal in the retrofitting from an
anoxic selector to an anaerobic selector in a full-scale activated
sludge process in Singapore. Water Science and Technology,
59(5): 857-865.
Cech, J.S. and Hartman, P. (1993). Competition between
polyphosphate and polysaccharide accumulating bacteria in
enhanced biological phosphate removal systems. Water
Research, 27(7): 1219-1225.
Chen, G.-H., Brdjanovic, D., Ekama, G.A. and van Loosdrecht,
M.C.M. (2010). Seawater as alternative water resource. In:
Proceedings of the 7th IWA Leading Edge Technology
Conference on Water and Wastewater Treatment, Arizona,
USA, June 2-4.
Chen, H., Yu, J.-J., Jia, X.-Y., and Jin, R.-C. (2014). Enhancement
of Anammox performance by Cu(II), Ni(II) and Fe(III)
supplementation. Chemosphere, 117 (1): 610-616.
Comeau, Y., Hall, K.J., Hancock, R.E.W. and Oldham, W.K.
(1986). Biochemical model for enhanced biological phosphorus
removal. Water Research, 20(12): 1511-1521.
Comeau, Y., Rabionwitz, B., Hall, K.J. and Oldham, W.K. (1987).
Phosphate release and uptake in enhanced biological
126
phosphorus removal from wastewater. Journal (Water
Pollution Control Federation), 59(7): 707-715.
Comeau Y. (2008). Microbial metabolism. In: Biological
wastewater treatment: principles, modelling and design, Henze,
M, van Loosdrecht, M.C.M., Ekama, G.A., Brdjanovic, D.
(Eds.), ISBN: 9781843391883, IWA Publishing. London, UK.
Corominas, L., Rieger, L., Takács, I., Ekama, A.G., Hauduc, H.,
Vanrolleghem, P., Oehmen, A., Gernaey, K., van Loosdrecht
M.C.M. and Comeau, Y. (2010). New framework for
standardized notation in wastewater treatment modelling.
Water Science and Technolology, 61(4): 841-857.
Cravo-Laureau, C., Matheron, R., Joulian, C., Cayol, J.-L. and
Hirschler-Réa, A. (2004). Desulfatibacillum alkenivorans sp.
nov., a novel nalkene-degrading, sulfate-reducing bacterium,
and emended description of the genus Desulfatibacillum.
International Journal of Systematic and Evolutionary
Microbiology, 54:1639-1642.
Cypionka. (1987). Uptake of sulfate, sulfite and thiosulfate by
proton-anion symport in Desulfovibrio desulfuricans. Archives
of Microbiology, 148(2): 144-149.
Cypionka, H. (2000). Oxygen respiration by Desulfovibrio species.
Annual Reviews of Microbiology, 54: 827-848.
Dai, Y., Yuan, Z., Wang, X., Oehmen, A. and Keller, J. (2007).
Anaerobic metabolism of Defluviicoccus vanus related
glycogen accumulating organisms (GAOs) with acetate and
propionate as carbon sources. Water Research, 41(9): 18851896.
Daigger, G.T., Hodgkinson, A., Aquilina, S. and Fries, M.K. (2015).
Development and implementation of a novel sulfur removal
process from H2S containing wastewaters. Water Environment
Research, 87(7): 618-625.
Dapena-Mora, A., Fernandez, I., Campos, J.L., Mosquera-Corral,
A., Mendez, R. and Jetten M.S.M. (2007). Evaluation of
activity and inhibition effects on Anammox process by batch
tests based on the nitrogen gas production. Enzyme and
Microbial Technology, 40(4): 859-865.
de Graaff, M.S., Temmink, H., Zeeman, G., van Loosdrecht,
M.C.M. and Buisman C.J.N. (2011). Autotrophic nitrogen
removal from black water: calcium addition as a requirement
for settleability. Water Research, 45: 63-74.
Dircks, K., Beun, J.J., van Loosdrecht, M.C.M., Heijnen, J.J. and
Henze, M. (2001). Glycogen metabolism in aerobic mixed
cultures. Biotechnology and Bioengineering, 73(2): 85-94.
Dircks, K., Henze, M., van Loosdrecht, M.C.M., Mosbaek, H. and
Aspegren, H. (2001). Storage and degradation of poly-Bhydroxybutyrate in activated sludge under aerobic conditions.
Water Research, 35: 2277-2285.
Dircks, K. Pind, P.F., Mosbaek, H. and Henze, M. (1999). Yield
determination by respirometry - The possible influence of
storage under aerobic conditions in activated sludge: Water SA,
25: 69-74.
Dosta, J., Fernandez, I., Vazquez-Padin, J. R., Mosquera-Corral, A.,
Campos, J.L., Mata-Alvarez, J. and Mendez, R. (2008). Shortand long-term effects of temperature on the Anammox
process. Journal of Hazardous Materials, 154(1): 688-693.
Eckenfelder, W. (1986). Operation control and management of
activated sludge plants treating industrial wastewaters.
Proceedings of a seminar sponsored by Vanderbilt University.
Tennessee.
Eikelboom, D.H. (2000). Process control of activated sludge plants
by microscopic investigation. ISBN-13: 9781780406831, IWA
Publishing, London, UK.
Ekama, G.A. and Wentzel, M.C. (2008a). Organic matter removal.
In: Biological wastewater treatment: principles, modelling and
design, Henze, M., van Loosdrecht, M.C.M., Ekama, G.A. and
Brdjanovic, D. (Eds.), ISBN: 9781843391883. IWA
Publishing, London, UK.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Ekama, G.A. and Wentzel, M.C. (2008b). Nitrogen removal. In:
Biological wastewater treatment: principles, modelling and
design, Henze, M., van Loosdrecht, M.C.M., Ekama, G.A. and
Brdjanovic, D. (Eds.), ISBN: 9781843391883, IWA
Publishing. London, UK.
Ekama, G.A., Wilsenach, J.A. and Chen, G.-H. (2010). Some
opportunities and challenges for urban wastewater treatment.
In: 7th IWA LET conference, Arizona, USA, June 2-4 (keynote
presentation), Retrieved in September 2015 from
http://repository.ust.hk/ir/Record/1783.1-16616
Ettwig, K.F., Butler, M.L., Le Paslier, D., Pelletier, E., Mangenot,
S., Kuypers, M.M.M., Schreiber, F., Dutilh, B.E., Zedelius, J.,
de Beer, D., Gloerich, J., Wessels, H.J.C.T., van Alen, T.,
Luesken, F., Wu, M.L., van de Pas-Schoonen, K.T., Op den
Camp, H.J.M., Janssen-Megens, E.M., Francoijs, K-J.,
Stunnenberg, H., Weissenbach, J., Jetten, M.S.M. and Strous,
M. (2010). Nitrite-driven anaerobic methane oxidation by
oxygenic bacteria. Nature, 464, 543-548.
Ficara, E. and Canziani, R. (2007). Monitoring denitrification by
pH-static titration. Biotechnology and Bioengineering, 98(2):
368-377.
Ficara, E. and Rozzi, A. (2004). Coupling pH-static and DO-stat
titration to monitor degradation of organic substrates. Water
Science and Technology, 49(1): 69-77.
Ficara, E., Cortelezzi, P. and Rozzi, A. (2003). Theory of pH-stat
titration, Biotechnology and Bioengineering, 82: 28-37.
Ficara, E., Musumeci, A. and Rozzi, A. (2000). Comparison and
combination of titrimetric and respirometric techniques to
estimate nitrification kinetics parameters. Water SA, 26(2):
217-224.
Ficara, E., Sambusiti, C. and Canziani, R. (2009). Manometric
monitoring of biological denitrification. In: Proceeding of the
2nd IWA Specialized Conference in Nutrients Management in
Wastewater Treatment Processes, Krakow, Poland, 6-9
September, 2009. Lemtech Konsulting (Ed.), Krakow. pp: 6168.
Filipe, C.D., Daigger, G.T. and Grady, Jr, C.P. (2001a). A metabolic
model for acetate uptake under anaerobic conditions by
glycogen accumulating organisms: stoichiometry, kinetics, and
the effect of pH. Biotechnology and Bioengineering, 76(1): 1731.
Filipe, C.D., Daigger, G.T. and Grady, Jr, C.L. (2001b). pH as a key
factor in the competition between glycogen-accumulating
organisms and phosphorus-accumulating organisms. Water
Environment Research, 73(2): 223-232.
Flemming, H.C., Wingender, J., Mayer, C., Korstgens, V. and
Borchard, W. (2000). Cohesiveness in biofilm matrix polymers.
In: Community structure and cooperation of biofilms, Allison,
D., Gilbert, P., Lappin-Scott, H.M. and Wilson, M. (Eds.),
SGM Symposium Series, 59, Cambridge University Press,
Cambridge, UK, pp: 87-105.
Flowers, J.J., He, S., Carvalho, G., Brook, Peterson, S., López, C.,
Yilmaz, S., Zilles, J.L., Morgenroth, E., Lemos, P., Reis,
M.A.M., Crespo, M.T.B., Noguera, D.R. and McMahon, K.D.
(2008). Ecological differentiation of Accumulibacter in EBPR
reactors. In: Proceedings of the Water Environment
Federation, WEFTEC 2008, pp: 31-42.
Foxon, K.M., Brouckaert, C.J., Buckley, C.A. and Rozzi, A. (2002).
Denitrifying activity measurements using an anoxic titration
(pHstat) bioassay. Water Scence and Technology, 46(9): 211218.
Gayon, U. and Dupetit, G. (1883). La fermentation des nitrates.
Mem. Soc. Sci. Phys. Nat., Bordeaux 2(5): 35-36.
Gernaey, K., Bogaert, H., Massone, A., Vanrolleghem, P. and
Verstraete, W. (1997). On-line nitrification monitoring in
activated sludge with a titrimetric sensor. Environmental
Science and Technology, 31: 2350-2355.
ACTIVATED SLUDGE ACTIVITY TESTS
Gernaey, K., Bogaert, H., van Rolleghem, P., Massone, A., Rozzi,
A. and Verstraete, W. (1998). A titration technique for online
nitrification monitoring in activated sludge. Water Science and
Technology, 37(12): 103-110.
Gibson, G.R. (1990). Physiology and ecology of the sulphatereducing bacteria. Journal of Applied Bacteriology, 69(6): 769797.
Goel, R., Mino, T., Satoh, H. and Matsuo, T. (1999). Modelling
hydrolysis processes considering intracellular storage: Water
Science and Technology, 39: 97-105.
Grady, Jr, L.C.P., Daigger, G.T., Love. N.G., and Filipe, C.D.M.
(2011). Biological wastewater treatment. 3rd edition, ISBN
9780849396793, IWA Publishing, CRC Press, London, UK.
Grein, F., Ramos, A.R., Venceslau, S.S. and Pereira, I.A.C. (2013).
Unifying concepts in anaerobic respiration: Insights from
dissimilatory sulfur metabolism. Biochimica et Biophysica Acta
(BBA) - Bioenergetics, 1827(2): 145-160.
Guisasola, A., Qurie, M., Vargas, M, Casas, C. and Baeza, J.A.
(2009). Failure of an enriched nitrite-DPAO population to use
nitrate as an electron acceptor. Process Biochemistry, 44: 689695.
Guisasola, G., Pijuan, M., Baeza, J.A., Carrera, J., Casas, C. and
Lafuente, J. (2004). Aerobic phosphorus release linked to
acetate uptake in bio-P sludge: process modelling using oxygen
uptake rate. Biotechnology and Bioengineering, 85: 721-733.
Guisasola, A., Sin, G., Baeza, J.A., Carrera, J. and Vanrolleghem,
P.A. (2005). Limitations of ASM 1 and ASM 3: a comparison
based on batch oxygen uptake rate profiles from different fullscale wastewater treatment plants. Water Science and
Technology, 52(10): 69-77.
Gujer, W., Henze, M., Mino, T. and van Loosdrecht, M.C.M.
(1999). Activated sludge model No. 3. Water Science and
Technology, 39(1): 183-193.
Hallin, S., Throback, I.N., Dicksved, J. and Pell, M. (2006).
Metabolic profiles and genetic diversity of denitrifying
communities in activated sludge after addition of methanol or
ethanol. American Society for Microbiology, 72(8): 5445-5452.
Hansen, T.A. (1993). Carbon metabolism of sulfate-reducing
bacteria. In: The sulfate-reducing bacteria: contemporary
perspectives, Odom, J.M., Singleton, Jr, R. (Eds.), ISBN 9781-4613-9263-7, Brock/Springer Book Series in Contemporary
Bioscience, Springer-Verlag, New York Inc.
Hansen, T.A. (1994). Metabolism of sulfate-reducing prokaryotes.
Antonie van Leeuwenhoek, 66(1-3): 165-185.
Henze, M. (1991). Capabilities of biological nitrogen removal
processes from wastewater. Water Science and Techology,
23(4-6): 669-679.
Henze, M. (1992). Characterization of wastewater for modeling of
activated sludge processes. Water Science and Technology,
25(6): 1-15.
Henze, M., Harremoes, P., Jansen, J.L.C. and Arvin, E. (1997).
Wastewater treatment: biological and chemical processes.
ISBN 978-3-540-42228-0, Springer, Berlin.
Henze, M. and Comeau, Y. (2008). Wastewater characterization. In:
Biological wastewater treatment: principles, modelling and
design, Henze, M., van Loosdrecht, M.C.M., Ekama, G.A. and
Brdjanovic, D. (Eds.), ISBN: 9781843391883, IWA
Publishing, London, UK.
Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais,
G.v.R. and van Loosdrecht, M.C.M. (1999). Activated sludge
model no. 2d, ASM2d. Water Science and Technology, 39(1):
165-182.
Henze, M., Gujer, W., Mino, T. and van Loosdrecht, M.C.M.
(2000). Activated sludge models ASM1, ASM2, ASM2d and
ASM3. IWA Scientific and Technical Report No. 9, IWA
Publishing, London, UK.
Henze, M., van Loosdrecht, M.C.M., Ekama, G.A. and Brdjanovic,
D. (2008). Biological wastewater treatment: principles,
127
modelling and design. ISBN: 9781843391883, IWA
Publishing, London, UK.
Hoover, S.R. and Porges, N. (1952). Assimilation of dairy wastes
by activated sludge II: The equation of synthesis and oxygen
utilization. Sewage and Industrial wastes, 24(3): 306-312.
Hu, Z., Lotti, T., van Loosdrecht, M.C.M. and Kartal, B. (2013).
Nitrogen removal with the anaerobic ammonium oxidation
process. Biotechnology Letters, 35(8):1145-1154.
Hu, J.Y., Ong, S.L., Ng, W.J., Lu, F. and Fan, X.J. (2003). A new
method for characterizing denitrifying phosphorus removal
bacteria by using three different types of electron
acceptors. Water Research, 37(14): 3463-3471.
Hu, Z.R., Wentzel, M.C. and Ekama, G.A. (2002). Anoxic growth of
phosphate-accumulating organisms (PAOs) in biological nutrient
removal activated sludge systems. Water Research, 36(19):
4927-4937.
Hulsbeek, J.J.W., Kruit, J., Roeleveld, P.J. and van Loosdrecht,
M.C.M. (2002). A practical protocol for dynamic modelling of
activated sludge systems. Water Science and Technology,
45(6): 127-136.
Hulshoff Pol, L.W., Lens, P.N.L, Stams, A.J.M. and Lettinga, G.
(1998). Anaerobic treatment of sulphate-rich wastewaters.
Biodegradation, 9(3-4): 213-224.
Isaksen, M.F. and Jørgensen, B.B. (1996). Adaptation of psychrophilic
and psychrotrophic sulfate-reducing bacteria to permanently cold
marine environments. Applied Environmental Microbiology, 62:
408-414.
Jobbagy, A., Literathy, B., Wong, M., Tardy, G. and Liu, W. (2006).
Proliferation of glycogen accumulating organisms induced by
Fe (III) dosing in a domestic wastewater treatment plant. Water
Science and Technology, 54(1): 101-109.
Jones, R., Dold, P., Takács, I., Chapman, K., Wett, B., Murthy, S.
and Shaughnessy, M. (2007). Simulation for operation and
control of reject water treatment processes. In: Proceedings of
the Water Environment Federation WEFTEC 2007, pp: 43574372.
Jørgensen, B.B. (2006). Bacteria and marine biogeochemistry. In:
Marine geochemistry, 2nd edition, Schulz, H.D. and Zabel, M.
(Eds.), Springer-Verlag, Berlin Heidelberg, pp: 169-206.
Jubany, I., Carrera, J., Lafuente, J. and Baeza, J.A. (2008). Start-up
of a nitrification system with automatic control to treat highly
concentrated ammonium wastewater: Experimental results and
modeling. Chemical Engineering Journal, 144(3): 407-419.
Kampschreur, M.J., Picioreanu, C., Tan, N.C.G., Kleerebezem, R.,
Jetten, M.S.M. and van Loosdrecht, M.C.M. (2007). Unraveling
the source of nitric oxide emission during nitrification. Water
Environment Research, 79: 2499-2509.
Kamyshny, Jr., A., Zilberbrand, M., Ekeltchik, I., Voitsekovski, T.,
Gun, J. and Lev, O. (2008). Speciation of polysulfides and zerovalent sulfur in sulfide-rich water wells in southern and central
Israel. Aquatic Geochemistry, 14: 171-192.
Kappeler, J. and Gujer, W. (1992). Estimation of kinetic parameters
of heterotrophic biomass under aerobic conditions and
characterization of wastewater for activated sludge modelling.
Water Science and Technology, 25(6): 125-139.
Kartal, B., Maalcke, W.J., de Almeida, N.M., Cirpus, I., Gloerich,
J., Geerts, W., den Camp, H.J.M.O., Harhangi, H.R., JanssenMegens, E.M., Francoijs, K.-J., Stunnenberg, H.G., Keltjens
J.T., Jetten, M.S.M. and Strousm, M. (2011). Molecular
mechanism of anaerobic ammonium oxidation. Nature, 479:
127-130.
Kerrn-Jespersen, J.P. and Henze, M. (1993). Biological phosphorus
uptake under anoxic and aerobic conditions. Water Research, 27:
617-624
Khanal, S.K. (2008). Anaerobic Biotechnology for Bioenergy
Production: Principles and Applications, Wiley-Blackwell,
ISBN: 978-0-8138-2346-1, Iowa, USA.
128
Kjeldsen, K.U., Joulian, C. and Ingvorsen, K. (2004). Oxygen
tolerance of sulfate-reducing bacteria in activated sludge.
Environmental Science & Technology, 38(7): 2038-2043.
Kjeldsen, K.U., Loy, A., Jakobsen, T.F., Thomsen, T.R., Wagner,
M. and Ingvorsen, K. (2007). Diversity of sulfate-reducing
bacteria from an extreme hypersaline sediment, Great Salt Lake
(Utah). FEMS Microbiology Ecology, 60: 287-298.
Koch, G., Egli, K., van der Meer, J.R. and Siegrist, H. (2000).
Mathematical modeling of autotrophic denitrification in a
nitrifying biofilm of a rotating biological contactor. Water
Science and Technology, 41(4-5): 191-198.
Koch G., Kuhni M., Gujer W. and Siegrist H. (2000). Calibration
and validation of activated sludge model No. 3 for Swiss
municipal wastewater. Water Research, 34: 3580-3590.
Kong, Y., Nielsen, J.L. and Nielsen, P.H. (2005). Identity and
ecophysiology of uncultured actinobacterial polyphosphateaccumulating organisms in full-scale enhanced biological
phosphorus removal plants. Applied and Environmental
Microbiology, 71(7): 4076-4085.
Kong, Y., Xia, Y., Nielsen, J.L. and Nielsen, P.H., (2006).
Ecophysiology of a group of uncultured Gammaproteobacterial
glycogen-accumulating organisms in full-scale enhanced
biological phosphorus removal wastewater treatment plants.
Environmental Microbiolology, 8(3): 479-489.
Könneke, M., Bernhard, A.E., de la Torre, J.R., Walker, C.B.,
Waterbury, J.B. and Stahl, D.A. (2005). Isolation of an
autotrophic ammonia-oxidizing marine archaeon. Nature,
437(7058): 543-6.
Kuba, T., Smolders, G., van Loosdrecht, M.C.M. and Heijnen, J.J.
(1993). Biological phosphorus removal from wastewater by
anaerobic-anoxic sequencing batch reactor. Water Science and
Technology, 27(5-6): 241-252.
Kuba, T., Murnleitner, E., van Loosdrecht, M.C.M. and Heijnen, J.J.
(1996). A metabolic model for biological phosphorus removal
by denitrifying organisms. Biotechnology and Bioengineering,
52: 685-695.
Kuba, T., van Loosdrecht, M.C.M., Brandse, F.A. and Heijnen, J.J.
(1997a). Occurrence of denitrifying phosphorus removing
bacteria in modified UCT-type wastewater treatment
plants. Water Research, 31(4): 777-786.
Kuba, T., van Loosdrecht, M.C.M., Murnleitner, E. and Heijnen, J.J.
(1997b). Kinetics and stoichiometry in the biological
phosphorus removal process with short cycle times. Water
Research, 31(4): 918-928.
Kujawa, K. and Klapwijk, B. (1999). A method to estimate
denitrification potential for pre-denitrification systems using
NUR batch tests. Water Research, 33: 2291-2300.
Kuypers, M.M.M., Lavik, G., Woebken, D., Schmid, M., Fuchs,
B.M., Amann, R., Jørgensen, B.B. and Jetten, M.S.M. (2005).
Massive nitrogen loss from the Benguela upwelling system
through anaerobic ammonium oxidation. Proceedings of the
National Academy of Sciences of the United States of America,
102: 6478-6483.
Lackner, S., Gilbert, E.M., Vlaeminck, S.E., Joss, A., Horn, H.
and van Loosdrecht, M.C.M. (2014). Full-scale partial
Nitritation/Anammox experiences - an application survey.
Water Research, 55: 292-303.
Langergraber, G., Rieger, L., Winkler, S., Alex, J., Wiese, J.,
Owerdieck, C., Ahnert, M., Simon, J. and Maurer. M. (2004).
A guideline for simulation studies of wastewater treatment
plants. Water Science and Technology, 50(7): 131-138.
Lanham, A.N., Ricardo, A.R., Coma, M., Fradinho, J., Carvalheira,
M., Oehmen, A., Carvalho, G. and Reis, M.A.M. (2012).
Optimisation of glycogen quantification in mixed microbial
cultures. Bioresource Technology, 118: 518-525.
Lanham, A.B., Oehmen, A., Saunders, A.M., Carvalho, G., Nielsen,
P.H. and Reis, M.A. (2014). Metabolic modelling of full-scale
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
enhanced biological phosphorus removal sludge. Water
Research, 66: 283-295.
Lau, G.N., Sharma, K.R., Chen, G.-H. and van Loosdrecht, M.C.M.
(2006). Integration of sulfate reduction autotrophic
denitrification and nitrification to achieve low-cost sludge
minimization for Hong Kong sewage. Water Science and
Technology, 53(3): 227-235.
Lens, P.N., De Poorter, M.-P., Cronenberg, C.C. and Verstraete,
W.H. (1995). Sulfate reducing and methane producing bacteria
in aerobic wastewater treatment systems. Water Research,
29(3): 871-880.
Lens, P.N.L., Visser, A., Janssen, A.J.H., Hulshoff Pol, L.W. and
Lettinga, G. (1998). Biotechnological treatment of sulfate-rich
wastewaters. Critical Reviews in Environmental Science and
Technology, 28(1): 41-88.
Li, L., Han, Y., Yan, X. and Liu, J. (2013). H2S removal and
bacterial structure along a full-scale biofilter bed packed with
polyurethane foam in a landfill site. Bioresource Technology,
147: 52-58.
Liamleam, W. and Annachhatre, A.P. (2007). Electron donors for
biological sulfate reduction. Biotechnology Advances, 25(5):
452-463.
Lin, Y.M., Bassin, J.P. and van Loosdrecht, M.C.M. (2012). The
contribution of exopolysaccharides induced struvites
accumulation to ammonium adsorption in aerobic granular
sludge. Water Research, 46(4): 986-992.
Little, B.J. and Lee, J.S. (2007). Microbiologically influenced
corrosion. ISBN 978-0-471-77276-7, John Wiley & Sons Inc.,
Hoboken, NJ, USA.
Liu, M.C and Peck, Jr., H.D. (1981). The isolation of a hexaheme
cytochrome from Desulfovibrio desulfuricans and its
identification as a new type of nitrite reductase. Journal of
Biological Chemistry, 256(24): 13159-13164.
Lochtman, S.F.W. (1995). Proceskeuze en -optimalisatie van het
SHARON proces voor slibverwerkingsbedrijf Sluisjesdijk
(Process choice and optimisation of the SHARON process for
the sludge treatment plant Sluisjesdijk). BODL report: TU
Delft.
Lopez, C., Pons, M.N. and Morgenroth, E. (2006). Endogenous
processes during long-term starvation in activated sludge
performing enhanced biological phosphorus removal. Water
Research, 40(8): 1519-1530.
Lopez-Vazquez, C.M., Song, Y.I., Hooijmans, C.M., Brdjanovic,
D., Moussa, M.S., Gijzen, H.J. and van Loosdrecht, M.C.M.
(2007). Short term temperature effects on the anaerobic
metabolism
of
glycogen
accumulating
organisms.
Biotechnology and Bioengineering, 97(3): 483-495.
Lopez-Vazquez, C.M., Hooijmans, C.M., Brdjanovic, D., Gijzen,
H.J. and van Loosdrecht, M.C.M. (2008a). Factors affecting the
microbial populations at full-scale enhanced biological
phosphorus removal (EBPR) wastewater treatment plants in
The Netherlands. Water Research, 42(10): 2349-2360.
Lopez-Vazquez, C.M., Brdjanovic, D. and van Loosdrecht, M.C.M.
(2008b). Comment on “Could polyphosphate-accumulating
organisms (PAOs) be glycogen-accumulating organisms
(GAOs)?” by Zhou, Y., Pijuan, M., Zeng, R.J, Lu, H. and Yuan.
Z. Water Research (2008). doi: 10.1016/j. waterres. 2008.01.
003, Water Research, 42(13): 3561-3562.
Lopez-Vazquez, C.M. (2009). The competition between
polyphosphate-accumulating organisms and glycogenaccumulating organisms: temperature effects and modelling.
PhD Thesis. ISBN 9780415558969, Delft University of
Technology, CRC Press/Balkema, Leiden, the Netherlands.
Lopez-Vazquez, C.M., Hooijmans, C.M., Brdjanovic, D., Gijzen,
H.J. and van Loosdrecht, M.C.M. (2009a). Temperature effects
on glycogen accumulating organisms. Water Research, 43(11):
2852-2864.
ACTIVATED SLUDGE ACTIVITY TESTS
Lopez-Vazquez, C.M., Oehmen, A., Hooijmans, C.M., Brdjanovic,
D., Gijzen, H.J., Yuan, Z. and van Loosdrecht, M.C.M.
(2009b). Modeling the PAO–-GAO competition: effects of
carbon source, pH and temperature. Water Research, 43(2):
450-462.
Lopez-Vazquez, C.M., Kubare, M., Saroj, D.P., Chikamba, C.,
Schwarz, J., Daims, H. and Brdjanovic, D. (2014). Thermophilic
biological nitrogen removal in industrial wastewater treatment.
Applied Microbiology and Biotechnology, 98(2): 945-956.
Lotti, T., van der Star, W.R., Kleerebezem, R., Lubello, C. and van
Loosdrecht, M.C.M. (2012). The effect of nitrite inhibition on
the anammox process. Water Research, 46(8): 2559-2269.
Lotti, T., Kleerebezem, R., Lubello, C. and van Loosdrecht, M.C.M.
(2014). Physiological and kinetic characterization of a suspended
cell Anammox culture. Water Research, 60(14): 1-14.
Lotti, T., Kleerebezem, R., Hu, Z., Kartal, B., de Kreuk, M.K., van
Erp Taalman Kip, C., Kruit, J., Hendrickx, T.L.G. and van
Loosdrecht, M.C.M. (2015a). Pilot-scale evaluation of anammox
based main-stream nitrogen removal from municipal wastewater.
Environmental Technology, 36(9): 1167-1177.
Lotti, T., Kleerebezem, R., Abelleira-Pereira, J.M., Abbas, B. and
van Loosdrecht M.C.M. (2015b). Faster through training: the
Anammox case. Water Research, 81: 261-268.
Lotti, T., Kleerebezem, R. and van Loosdrecht, M.C.M. (2015c).
Effect of temperature change on Anammox activity.
Biotechnology and Bioengineering, 112(1): 98-103.
Madigan, M.T., Martinko, J.M., Dunlap, P.V. and Clark, D.P.
(2009). Brock biology of microorganisms, 12th edition, ISBN
0-13-232460-1, Pearson Benjamin Cummings, San Francisco,
USA.
Mañas, A., Biscans, B. and Spérandio, M. (2011). Biologically
induced phosphorus precipitation in aerobic granular sludge
process. Water Research, 45(12): 3776-3786.
Marais, G.v.R. and Ekama, G.A. (1976). The activated sludge
process part 1 - Steady state behaviour. Water SA, 2(4): 163200.
Martins, A.M.P. (2004). Bulking sludge control: kinetics, substrate
storage, and process design aspects. PhD Thesis. ISBN 9729098-07-7, Delft University of Technology, Delft, The
Netherlands.
Martins, A.M.P., Heijnen, J.J. and van Loosdrecht, M.C.M. (2003).
Effect of feeding pattern and storage on sludge settleability
under aerobic conditions. Water Research, 37(11): 2555-2570.
Massone, A., Antonelli, M. and Rozzi, A. (1996). The DENICON:
a novel biosensor to control denitrification in biological
wastewater treatment plants. Mededelingen Faculteit
Landbouwkundige, University of Gent, 1709-1714.
Massone, A.G., Gernaey, K., Rozzi, A. and Verstraete, W. (1998).
Measurement of ammonium concentration and nitrification rate
by a new titrometric biosensor. WEF Research Journal, 70(3):
343-350.
Mateju, V., Cizinska, S., Krejci, J. and Janoch, T. (1992). Biological
water denitrification - a review. Enzyme and Microbial
Technology, 14: 170-183.
McCarty, P.L. (2007). Thermodynamic electron equivalents model
for bacterial yield prediction: Modifications and comparative
evaluations. Biotechnology and Bioengineering, 97: 377-388.
Meijer, S.C.F. (2004). Theoretical and practical aspects of
modelling activated sludge processes. PhD Thesis. ISBN 909018027-3, Delft University of Technology, Delft, the
Netherlands.
Meinhold, J., Arnold, E. and Isaacs, S. (1999). Effect of nitrite on
anoxic phosphate uptake in biological phosphorus removal
activated sludge. Water Research, 33(8): 1871-1883.
Mesquita D.P., Amaral A.L., Ferreira E.C. (2013) Activated sludge
characterization through microscopy: a review on quantitative
image analysis and chemometric techniques. Analytica
Chimica Acta, 802: 14-28.
129
Metcalf and Eddy (2003). Wastewater engineering: treatment,
disposal and reuse. 4th edition, ISBN-13:978-0070418783,
McGraw-Hill, Boston, USA.
Mino, T., van Loosdrecht, M.C.M. and Heijnen, J.J. (1998).
Microbiology and biochemistry of the enhanced biological
phosphate removal process. Water Research, 32(11): 3193-3207.
Mino, T., Arun, V., Tsuzuki, Y. and Matsuo, T. (1987). Effect of
phosphorus accumulation on acetate metabolism in the
biological phosphorus removal process. In: biological
phosphate removal from wastewaters, Ramadori, R. (Ed.),
Pergamon Press, Oxford, United Kingdom. pp: 27-38.
Mohan, S.V., Rao, N.C., Prasad, K.K. and Sarma, P.N. (2005).
Bioaugmentation of an anaerobic sequencing batch biofilm
reactor (AnSBBR) with immobilized sulphate reducing
bacteria (SRB) for the treatment of sulphate bearing chemical
wastewater. Process Biochemistry, 40(8): 2849-2857.
Mokhayeri, Y., Nichols, A., Murthy, S., Riffat, R., Dold, P. and
Takács, I. (2006). Examining the influence of substrates and
temperature on maximum specific growth rate of denitrifiers.
Water Science and Technology, 54(8): 155-162.
Mokhayeri, Y., Riffat, R., Takács, I., Dold, P., Bott, C., Hinojosa,
Bailey, W. and Murthy S. (2008). Characterizing denitrification
kinetics at cold temperature using various carbon sources in labscale sequencing batch reactors. Water Science and Technology,
58(1): 233-238.
Mori, K., Kim, H., Kakegawa, T. and Hanada, S. (2003). A novel
lineage of sulfate-reducing microorganisms: Thermodesulfobiaceae
fam. nov., Thermodesulfobium narugense, gen. nov., sp. nov., a new
thermophilic isolate from a hot spring. Extremophiles, 7: 283290.
Moussa, M.S., Fuentes, O.G., Lubberding, H.J., Hooijmans, C.M.,
van Loosdrecht, M.C.M. and Gijzen, H.J. (2006). Nitrification
activities in full-scale treatment plants with varying salt loads.
Environmental Technology, 27(6): 635-643.
Moussa, M.S., Rojas, A.R., Hooijmans, C.M., Gijzen, H.J. and van
Loosdrecht, M.C.M. (2004). Model-based evaluation of
nitrogen removal in a tannery wastewater treatment plant.
Water Science and Technology, 50(6): 251-60.
Moussard, H., L’Haridon, S., Tindall, B.J., Banta, A., Schumann,
P., Stackebrandt, E., Reysenbach, A.-L., and Jeanthon, C.
(2004). Thermodesulfatator indicus gen. nov., sp. nov., a novel
thermophilic
chemolithoautotrophic
sulfate-reducing
bacterium isolated from the Central Indian Ridge. International
Journal of Systematic and Evolutionary Microbiology, 54: 227233.
Mulder, A. (2003). The quest for sustainable nitrogen removal
technologies. Water Science and Technology, 48(1): 67-75.
Munro, J.H.M. (1886). The formation and destruction of nitrates
and nitrates in artificial solutions and in river and well waters.
Journal of the Chemical Society, Transactions, 49: 632-681.
Murnleitner, E., Kuba, T., van Loosdrecht, M.C.M. and Heijnen, J.J.
(1997). An integrated metabolic model for the aerobic and
denitrifying biological phosphorus removal. Biotechnology and
Bioengineering, 54(5): 434-450.
Muyzer, G. and Stams, A.J.M. (2008). The ecology and biotechnology
of sulphate-reducing bacteria. Nature, 6:441-455.
Naidoo, V., Urbain, V. and Buckley, C.A. (1998). Characterisation
of wastewater and activated sludge from European municipal
wastewater treatment plants using the NUR test. Water Science
and Technology, 38(1): 303-310.
Nelson, D.L. and Cox, M.M. (2005). Lehninger principles of
biochemistry. ISBN-10: 1-4292-33414-8, W.H. Freeman and
Company, New York, USA.
Nielsen, P.H. and Keiding K. (1998). Disintegration of activated
sludge flocs in presence of sulfide. Water Research, 32(2): 313320.
Nielsen, P.H., Mielczarek, A.T., Kragelund, C., Nielsen, J.L.,
Saunders, A.M., Kong, Y., Hansen, A.A. and Vollertsen, J.
130
(2010). A conceptual ecosystem model of microbial
communities in enhanced biological phosphorus removal
plants. Water Research, 44(17): 5070-5088.
Oehmen, A., Yuan, Z., Blackall, L.L. and Keller, J. (2004). Short-term
effects of carbon source on the competition of polyphosphate
accumulating
organisms
and
glycogen
accumulating
organisms. Water Science and Technology, 50(10): 139-144.
Oehmen, A., Vives, M.T., Lu, H., Yuan, Z. and Keller, J. (2005a).
The effect of pH on the competition between polyphosphateaccumulating
organisms
and
glycogen-accumulating
organisms. Water Research, 39(15): 3727-3737.
Oehmen, A., Keller-Lehmann, B., Zeng, R.J., Yuan, Z. and Keller,
J. (2005b). Optimisation of poly-β-hydroxyalkanoate analysis
using gas chromatography for enhanced biological phosphorus
removal systems. Journal of Chromatography A, 1070(1-2):
131-136.
Oehmen, A., Yuan, Z., Blackall, L.L. and Keller, J. (2005c).
Comparison of acetate and propionate uptake by polyphosphate
accumulating organisms and glycogen accumulating
organisms. Biotechnology and Bioengineering, 91(2): 162-168.
Oehmen, A., Saunders, A.M., Vives, M.T., Yuan, Z. and Keller, J.
(2006). Competition between polyphosphate and glycogen
accumulating organisms in enhanced biological phosphorus
removal systems with acetate and propionate as carbon
sources. Journal of Biotechnology, 123(1): 22-32.
Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z., Keller, J.,
Blackall, L.L. and Reis, M.A. (2007). Advances in enhanced
biological phosphorus removal: from micro to macro
scale. Water Research, 41(11): 2271-2300.
Ong, Y.H., Chua, A.S.M., Fukushima, T., Ngoh, G.C., Shoji, T. and
Michinaka A. (2014). High-temperature EBPR process: The
performance, analysis of PAOs and GAOs and the fine-scale
population
study
of
Candidatus
“Accumulibacter
phosphatis”. Water Research, 64: 102-112.
Orhon, D., Sözen, S. and Artan, N. (1996). The effect of
heterotrophic yield on the assessment of the correction factor
for anoxic growth. Water Science and Technology, 34(5): 6774.
Oshiki, M., Shimokawa, M., Fujii, N., Satoh, H. and Okabe, S.
(2011). Physiological characteristics of the anaerobic
ammonium-oxidizing bacterium ‘Candidatus Brocadia sinica’.
Microbiology, 157: 1706-1713.
Oude Elferink, S.J.W.H., Visser, A., Hulshoff Pol, L.W. and Stams,
A.J.M. (1994). Sulfate reduction in methanogenic bioreactors.
FEMS Microbiology Reviews, 15: 119-136.
Parrou, J.L. and François, J. (1997). A simplified procedure for a
rapid and reliable assay of both glycogen and trehalose in whole
yeast cells. Analytical Biochemistry, 248(1): 186-188.
Pasteur, L. (1859). Note sur la fermentation nitreuse. Bulletin de la
Société de chimique de Paris (séance du 11 mars ): 22-23.
Payne, W.J. (1986). 1986: Centenary of the isolation of denitrifying
bacteria. ASM News 52(12): 627-629.
Petersen, B., Gernaey, K. and Vanrolleghem, P.A. (2002). Anoxic
activated sludge monitoring with combined nitrate and
titrimetric measurements. Water Science and Technology,
45(4-5): 181-190.
Pijuan, M., Guisasola, A., Baeza, J.A., Carrera, J., Casas, C. and
Lafuente, J. (2006). Net P-removal deterioration in enriched
PAO sludge subjected to permanent aerobic conditions.
Journal of Biotechnology, 123: 117-126.
Pijuan, M., Ye, L., and Yuan, Z. (2010). Free nitrous acid inhibition
on the aerobic metabolism of poly-phosphate accumulating
organisms. Water Research, 44(20): 6063-6072.
Pijuan, M., Ye, L. and Yuan, Z. (2011). Could nitrite/free nitrous acid
favour GAOs over PAOs in enhanced biological phosphorus
removal systems? Water Science and Technology, 63(2): 345351.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Pikuta, E., Lysenko, A., Chuvilskaya, N., Mendrock, U., Hippe, H.,
Suzina, N., Nikitin, D., Osipov, G. and Laurinavichius, K.
(2000). Anoxybacillus pushchinensis gen. nov., sp. nov., a
novel anaerobic, alkaliphilic, moderately thermophilic
bacterium from manure, and description of Anoxybacillus
flavitherms comb. Nov. International Journal of Systematic
and Evolutionary Microbiology, 50(6): 2109-2117.
Piña-Ochoa, E., Hogslund, S., Geslin, E. and Risgaard-Petersen, N.
(2010). Survival and life strategy of the foraminiferan
Globobulimina turgida through nitrate storage and
denitrification. Marine Ecology Progress Series, 417: 39-49.
Poinapen, J., Ekama, G.A. and Wentzel, M.C. (2009). Biological
sulphate reduction with primary sewage sludge in an upflow
anaerobic sludge bed (UASB) reactor - Part 2: Modification of
simple wet chemistry analytical procedures to achieve COD
and S mass balances. Water SA, 35(5): 535-542.
Postgate, J.R. (1965). Recent advances in the study of the sulfatereducing bacteria. Bacteriological Reviews, 29(4): 425-441.
Postgate, J.R. (1984). The Sulfate-Reducing Bacteria. 2nd Edition,
ISBN: 9780521257916, Cambridge University Press, UK.
Puyol, D., Carvajal-Arroyo, J.M., Garcia, B., Sierra-Alvarez, R. and
Field, J.A. (2013). Kinetic characterization of Brocadia spp.dominated Anammox cultures. Bioresource Technology, 139:
94-100.
Puyol, D., Carvajal-Arroyo, J.M., Sierra-Alvarez, R. and Field, J.A.
(2014). Nitrite (not free nitrous acid) is the main inhibitor of the
anammox process at common pH conditions. Biotechnology
Letters, 36(3): 547-551.
Rabus, R., Hansen, T.A. and Widdel, F. (2006). Dissimilatory
sulfate- and sulfur-reducing prokaryotes. In: The prokaryotes,
Vol. 2, Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.H. and Stackebrandt, E. (Eds.), New York: Springer, pp: 659768.
Ramdani, A., Dold, P., Gadbois, A., Deleris, S., Houweling D. and
Comeau, Y. (2012). Characterization of the heterotrophic
biomass and the endogenous residue of activated sludge. Water
Research, 46(3): 653-668.
Rebac, S., Visser, A., Gerbens, S., van Lier, J.B., Stams, A.J.M. and
Lettinga, G. (1996). The effect of sulphate on propionate and
butyrate degradation in a psychrophilic anaerobic expanded
granular sludge bed (EGSB) reactor. Environmental
Technology, 17(9): 997-1005.
Reichenbecher, W. and Schink. B. (1997). Desulfovibrio inopinatus
sp. nov., a new sulfate-reducing bacterium that degrades
hydroxyhydroquinone (1,2,4-trihydroxybenzene). Archives of
Microbiology, 168: 338-344.
Reiset, J. (1856). Experiences sur la putrefaction et sur la formation
des fumiers. Comptes rendus des séances hebdomadaires de
l’Académie des Sciences, 42: 177-180.
Rikmann, E., Zekker, I., Tomingas, M., Tenno, T., Menert, A.,
Loorits, L. and Tenno, T. (2012). Sulfate-reducing anaerobic
ammonium oxidation as a potential treatment method for high
nitrogen-content wastewater. Biodegradation, 23: 509-524.
Rintala, J.A. and Puhakka, J.A. (1994). Anaerobic treatment in pulp
and paper-mill waste management: A review. Bioresource
Technology, 47(1): 1-18.
Risgaard-Petersen, N., Langezaal, A.M., Ingvardsen, S., Schmid,
M. C., Jetten, M.S.M., Op den Camp, H.J.M., Derksen, J.W.M.,
Piña-Ochoa, E., Eriksson, S.P., Nielsen, L.P., Revsbech, N.P.,
Cedhagen, T. and van der Zwaan, G. J. (2006). Evidence for
complete
denitrification
in
a
benthic
foraminifer. Nature, 443(7107): 93-96.
Robertson, L.A., Dalsgaard, T., Revsbech, N.P. and Kuenen, J.G.
(1995). Confirmation of aerobic denitrification in batch
cultures, using gas-chromatography and N-15 massspectrometry. FEMS Microbiology Ecology, 18(2): 113-119.
ACTIVATED SLUDGE ACTIVITY TESTS
Roeleveld, P.J. and van Loosdrecht, M.C.M. (2002). Experience
with guidelines for wastewater characterisation in The
Netherlands. Water Science and Technology, 45(6): 77-87.
Rozzi, A., Castellazzi, L. and Speece, R.E. (2002). Acetoclastic
methanogenic activity measurements by a titration biosensor.
Biotechnology and Bioengineering, 77(1): 20-26.
Rozzi, A., Castellazzi, L. and Speece, R.E. (2002). Acetoclastic
methanogenic activity measurements by a titration biosensor.
Biotechnology and Bioengineering, 77(1): 20-26.
Rozzi, A., Ficara, E. and Rocco, A. (2003). DO-stat titration
respirometry: principle of operation and validation. ASCEJournal of Environmental Engineering, 129(7): 602-609.
Rozzi, A., Massone, A. and Antonelli, M. (1997). A VFA measuring
biosensor based on nitrate reduction. Water Science and
Technology, 36(6-7): 183-189.
Rubio-Rincon, F., Welles, L., López-Vázquez, C.M., van
Loosdrecht, M.C.M. and Brdjanovic, D., (2016). Sulfide effects
on the metabolism of Candidatus Accumulibacter phosphatis
clade I. (submitted).
Saad, S.A., Welles, L., López-Vázquez, C.M., van Loosdrecht,
M.C.M. and Brdjanovic. D. (2013). Sulfide effects on the
anaerobic kinetics of phosphorus-accumulating organisms. In:
Proceedings of 13th World Congress on Anaerobic Digestion,
25-28th June, Santiago de Compostela, Spain.
Saad, S.A., Welles, L., Abbas, B., López-Vázquez, C.M., van
Loosdrecht, M.C.M. and Brdjanovic, D. (2016). Denitrification
pathways of Candidatus Accumulibacter phosphatis clade I
using different carbon sources. (submitted).
Saito, T., Brdjanovic, D. and van Loosdrecht, M.C.M. (2004).
Effect of nitrite on phosphate uptake by phosphate
accumulating organisms. Water Research, 38(17): 3760-3768.
Sánchez, M., Mosquera-Corral, A., Mendez, R. and Lema, J.M.
(2000). Simple methods for the determination of the
denitrifying activity of sludges. Bioresource Technology, 75(16): 1-6.
Saunders, A.M., Oehmen, A., Blackall, L.L., Yuan, Z. and Keller,
J. (2003). The effect of GAOs on anaerobic carbon
requirements in full-scale Australian EBPR plants. Water
Science and Technology, 47(11): 37-43.
Scaglione, D., Buttiglieri, G., Ficara, E., Caffaz, S., Lubello, C. and
Malpei, F. (2009). Microcalorimetric and manometric tests to
assess anammox activity. Water Science and Technology,
60(10): 2705-2711.
Schuler, A.J. and Jenkins, D. (2003). Enhanced biological
phosphorus removal from wastewater by biomass with different
phosphorus contents, part I: experimental results and
comparison with metabolic models. Water Environment
Research, 75(6): 485-498.
Schlœsing, T. and Müntz, A. (1877). Sur la nitrification par les
ferments organics. Comptes rendus des séances hebdomadaires
de l’Académie des Sciences, 85: 301-303.
Sin, G., Guisasola, A., De Pauw, D.J.W., Baeza, J.A., Carrera, J.
and Vanrolleghem, P.A. (2003). A new approach for modelling
simultaneous storage and growth processes for activated sludge
systems under aerobic conditions. Biotechnology and
Bioengineering, 92(5): 600-613.
Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M. and
Heijnen, J.J. (1994a). Model of the anaerobic metabolism of the
biological phosphorus removal process: stoichiometry and pH
influence. Biotechnology and Bioengineering, 43(6): 461-470.
Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M. and
Heijnen, J.J. (1994b). Stoichiometric model of the aerobic
metabolism of the biological phosphorus removal
process. Biotechnology and Bioengineering, 44(7): 837-848.
Smolders, G.J.F., van Loosdrecht, M.C.M. and Heijnen, J.J. (1995). A
metabolic model for the biological phosphorus removal
process. Water Science and Technology, 31(2): 79-93.
131
Strous, M., Heijnen, J.J., Kuenen, J.G. and Jetten, M.S.M. (1998).
The sequencing batch reactor as a powerful tool for the study
of slowly growing anaerobic ammonium-oxidizing
microorganisms. Applied Microbiology and Biotechnology, 50:
589-596.
Tang, K., Baskaran, V. and Nemati, M. (2009). Bacteria of the
sulphur cycle: An overview of microbiology, biokinetics and
their role in petroleum and mining industries. Biochemical
Engineering Journal, 44(1): 73-94.
Tang, C.-J., Zheng, P., Chen, T.-T., Zhang, J.-Q., Mahmood, Q.,
Ding, S., Chen, X.-G., Chen, J.-W., and Wu D.-T. (2011).
Enhanced nitrogen removal from pharmaceutical wastewater
using SBA-ANAMMOX process. Water Research, 45: 201210.
Thauer, R.K., Jungermann, K. and Decker, K. (1977). Energy
conservation in chemotrophic anaerobic bacteria. Bacteriological
Reviews, 41: 100-180.
Thomas, M., Wright, P., Blackall, L.L., Urbain, V. and Keller, J.
(2003). Optimisation of Noosa BNR plant to improve
performance and reduce operating costs. Water Science and
Technology, 47(12): 141-148.
Udert, K.M., Kind, E., Teunissen, M., Jenni, S. and Larsen, T.A.
(2008). Effect of heterotrophic growth on nitritation/anammox
in a single sequencing batch reactor. Water Science and
Technology, 58: 277-284.
van de Graaf, A., De Bruijn, P., Robertson, L.A., Jetten, M.S.M. and
Kuenen, J.G. (1996). Autotrophic growth of anaerobic
ammonium oxidizing microorganisms in a fluidized bed
reactor. Microbiology, 142(8): 2187-2196.
van den Brand, T.P.H., Roest, K., Brdjanovic, D., Chen, G.-H. and
van Loosdrecht, M.C.M. (2014a). Influence of acetate and
propionate on sulphate-reducing bacteria activity. Journal of
Applied Microbiology, 117(6): 1839-1847.
van den Brand, T.P.H., Roest, K., Chen, G.-H., Brdjanovic, D. and
van Loosdrecht, M.C.M. (2014b). Temperature effect on
acetate and propionate consumption by sulphate reducing
bacteria in saline wastewater. Applied Microbiology and
Biotechnology, 98(9): 4245-4255.
van den Brand, T.P.H., Roest, K., Chen, G.-H., Brdjanovic, D. and
van Loosdrecht, M.C.M. (2015). Occurence and activity of
sulphate reducing bacteria in aerobic activated sludge systems.
World Journal of Microbiology and Biotechnology, 31(3): 507516.
van Houten, R.T., van der Spoel, H., van Aelst, A.C., Hulshoff Pol,
L.W. and Lettinga, G. (1996). Biological sulfate reduction
using synthesis gas as energy and carbon source. Biotechnology
and Bioengineering, 50(2): 136-144.
van Loosdrecht, M.C.M., Lopez-Vazquez, C.M., Meijer, S.C.F.,
Hooijmans, C.M., and Brdjanovic, D. (2015) Twenty-five years
of ASM1: past, present and future of wastewater treatment
modelling. Journal of Hydroinformatics, 17(5):697-718.
van Loosdrecht, M.C.M., Pot, M.A. and Heijnen, J.J. (1997).
Importance of bacterial storage polymers in bioprocesses.
Water Science and Technology, 35(1): 41-47.
van Niel, E.W.J., Arts, P.A.M., Wesselink, B.J., Robertson L.A. and
Kuenen, J.G. (1993). Competition between heterotrophic and
autotrophic nitrifiers for ammonia in chemostat cultures. FEMS
Microbiology Ecology, 102: 109-118.
Vanrolleghem, P.A., Insel, G., Petersen, B., Sin, G., De Pauw, D.,
Nopens, I., Doverman, H., Weijers, S. and Gernaey, K. (2003). A
comprehensive model calibration procedure for activated sludge
models. In: Proceedings of the 76th Annual WEF Conference and
Exposition, October 11-15, Los Angeles.
Villa-Gomez, D., Ababneh, H., Papirio, S., Rousseau, D.P.L. and
Lens, P.N.L. (2011). Effect of sulfide concentration on the
location of the metal precipitates in inversed fluidized bed
reactors. Journal of Hazardous Materials, 192(1): 200-207.
132
Villegas, J.D., de Laclos, H.F., Dovat, J., Membrez, Y. and Holliger,
C. (2011). Nitrogen removal from digested manure in a simple
one-stage process. Water Science and Technology, 63: 19911996.
Vishniac, W. and Santer, M. (1957). Thiobacilli. Bacteriological
Reviews, 21: 195-213.
Vlekke, G.J.F.M., Comeau, Y. and Oldham W.K. (1988). Biological
phosphate removal from wastewater with oxygen or nitrate in
sequencing batch reactors. Environmental Technology Letters,
9: 791-796.
Wachtmeister, A., Kuba, T., van Loosdrecht, M.C.M. and Heijnen, J.J.
(1997). A sludge characterization assay for aerobic and denitrifying
phosphorus removing sludge. Water Research, 31(3): 471-478.
Wang, J., Lu, H., Chen, G.H., Lau, G.N., Tsang, W.L. and van
Loosdrecht, M.C.M. (2009). A novel sulfate reduction,
autotrophic denitrification, nitrification integrated (SANI)
process for saline wastewater treatment. Water Research,
43(9): 2363-2372.
Welles, L., López-Vázquez, C.M., Hooijmans, C.M., van Loosdrecht,
M.C.M. and Brdjanovic, D. (2014). Impact of salinity on the
anaerobic metabolism of phosphate-accumulating organisms
(PAO) and glycogen-accumulating organisms (GAO). Applied
Microbiology and Biotechnology, 98(17): 7609-7622.
Welles, L., López-Vázquez, C.M., Hooijmans, C.M., van
Loosdrecht, M.C.M. and Brdjanovic, D. (2015a). Impact of
salinity on the aerobic metabolism of phosphate-accumulating
organisms. Applied Microbiology and Biotechnology, 99(8):
3659-3672.
Welles, L., Tian, W.D., Saad, S., Abbas, B., López-Vázquez, C.M.,
Hooijmans, C.M., van Loosdrecht, M.C.M. and Brdjanovic, D.
(2015b). Accumulibacter clades Type I and II performing
kinetically different glycogen-accumulating organisms
metabolisms for anaerobic substrate uptake. Water Research,
15(83): 354-366.
Welles, L., Abbas, B., López-Vázquez, C.M., Hooijmans, C.M., van
Loosdrecht, M.C.M. and Brdjanovic, D. (2016). Metabolic
response of 'Candidatus Accumulibacter phosphatis' clade II to
changes in P/C ratio in their environment. (submitted).
Wentzel, M C., Dold, P.L., Loewenthal, R.E., Ekama, G.A. and
Marais, G.v.R. (1987). Experiments towards establishing the
kinetics of biological excess phosphorus removal.
In: Biological phosphate removal from wastewaters:
Proceedings of an IAWPRC Specialized Conference, Rome,
Italy, 28-30 September 28-30, 1987 (Pergamon Press, Vol. 4,
p. 79)..
Wentzel, M.C., Comeau, Y., Ekama, G.A., van Loosdrecht, M.C.M.
and Brdjanovic, D. (2008). Enhanced biological phosphorus
removal. In: Biological wastewater treatment: principles,
modelling and design, Henze, M., van Loosdrecht, M.C.M.,
Ekama, G.A. and Brdjanovic, D. (Eds.), ISBN:
9781843391883, IWA Publishing, London, UK.
Wentzel, M.C., Lötter, L.H., Loewenthal, R.E and Marais, G.v.R.
(1986). Metabolic behaviour of Acinetobacter spp. in enhanced
biological phosphorus removal- a biochemical model. Water
SA, 12(4): 209-224.
Wentzel, M.C., Mbewe, A. and Ekama, G.A. (1995). Batch tests for
measurement of readily biodegradable COD and active
organism concentrations in municipal waste waters. Water SA,
21(2): 117-124.
WERF. (2003). Methods for wastewater characterization in
activated sludge modeling. - Water Environment Research
Foundation report 99-WWF-3, ISBN13: 9781843396628.
WERF (Alexandria) and IWA Publishing (London), pp. 575
Wett, B. (2007). Development and implementation of a robust
deammonification process. Water Science and Technology,
56(7): 81-88.
Widdel, F. (2006). The genus Desulfotomaculum. In: The
prokaryotes, Vol. 2, Dworkin, M., Falkow, S., Rosenberg, E.,
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Schleifer, K.-H., and Stackebrandt, E. (Eds.), New York:
Springer, pp: 787-794.
Wiesmann, U. (1994). Biological nitrogen removal from wastewater.
In: Advances in biochemical engineering/biotechnology,
Fiechter, A. (Ed.), ISBN-13: 978-1843396628, Berlin: SpringerVerlag Berlin Heidelberg. pp: 113-154.
Wik, T. and Breitholtz, C. (1996). Steady-state solution of a twospecies biofilm problem. Biotechnology and Bioengineering,
50(6): 675-686.
Winogradsky, M.S. (1890). Reserches sur les organismes de la
nitrification. Annales de l’institut Pasteur, 4: 213-231.
Winogradsky, M.S. (1892). Contribution à la morphologie des
organismes de la nitrification. Archives of Biological Sciences,
1: 87-137.
Wu, D., Ekama, G.A., Wang, H.G., Wei, L., Lu, H., Chui, H.K.,
Liu, W.T., Brdjanovic, D., van Loosdrecht, M.C.M. and Chen,
G.H. (2014). Simultaneous nitrogen and phosphorus removal in
the sulphur cycle-associated Enhanced Biological Phosphorus
Removal (EBPR) process. Water Research, 49: 251-264.
Yamamoto, R.I., Komori, T. and Matsui, S. (1991). Filamentous
bulking and hindrance of phosphate removal due to sulfate
reduction in activated sludge. Water Science and Technology,
23(4-6): 927-935.
Yang, Z., Zhou, S. and Sun, Y. (2009). Start-up of simultaneous
removal of ammonium and sulfate from an anaerobic
ammonium oxidation (Anammox) process in an anaerobic upflow bioreactor. Journal of Hazardous Materials, 169: 113118.
Yoshida. Y., Takahashi. K., Saito. T. and Tanaka. K. (2006). The
effect of nitrite on aerobic phosphate uptake and denitrifying
activity of phosphate-accumulating organisms. Water Science
and Technology, 53(6): 21-27.
Zeng. R.J., van Loosdrecht. M.C.M., Yuan. Z. and Keller. J. (2003a).
Metabolic model for glycogen-accumulating organisms in
anaerobic/aerobic activated sludge systems. Biotechnology and
Bioengineering, 81(1): 92-105.
Zeng. R.J., Saunders. A.M., Yuan. Z., Blackall. L.L. and Keller, J.
(2003b). Identification and comparison of aerobic and
denitrifying
polyphosphate-accumulating
organisms.
Biotechnology and Bioengineering, 83(2): 140-148.
Zeng, W., Li, B., Yang, Y., Wang, X., Li, L. and Peng, Y. (2014).
Impact of nitrite on aerobic phosphorus uptake by polyphosphate accumulating organisms in enhanced biological
phosphorus removal sludges. Bioprocess and Biosystems
Engineering, 37(2): 277-287.
Zhou, Y., Pijuan, M. and Yuan, Z. (2007). Free nitrous acid
inhibition on anoxic phosphorus uptake and denitrification by
polyphosphate accumulating organisms. Biotechnology and
Bioengineering, 98(4): 903-912.
Zhou, Y., Pijuan, M., Zeng, R.J., Lu, H. and Yuan, Z. (2008). Could
polyphosphate-accumulating organisms (PAOs) be glycogenaccumulating organisms (GAOs)? Water Research, 42(10):
2361-2368.
Zhou, Y., Ganda, L., Lim, M., Yuan, Z. and Ng, W.J. (2012).
Response of poly-phosphate accumulating organisms to free
nitrous acid inhibition under anoxic and aerobic
conditions. Bioresource Technology, 116: 340-347.
Zietz, U. (1995). The formation of sludge bulking in the activated
sludge process. European Water Pollution Control, 5: 21-27.
Zumft, W.G. (1997). Cell biology and molecular basis of
denitrification. Microbiology and Molecular Biology Reviews
61: 533-616.
3
RESPIROMETRY
Authors:
Reviewers:
Henri Spanjers
Peter A. Vanrolleghem
George A. Ekama
Mathieu Spérandio
3.1 INTRODUCTION
The objective of this chapter is to provide practical
guidelines for the assessment of the respiration rate of
biomass. The approach will be practically oriented and
method-driven. However, some biochemical background
on respiration will be provided in order to understand
how respiration is related to microbial substrate
utilization and growth. We will explain that respiration
can be assessed in terms of the uptake rate of a terminal
electron acceptor, such as molecular oxygen or nitrate,
or, in the case of anaerobic respiration, in terms of the
production rate of methane or sulphide. The
measurement of the consumption (or production) rate, i.e.
respirometry, will be explained following various
measuring principles, and we will provide some practical
recommendations. The focus will be on laboratory tests
using samples of biomass and wastewater. However, in
principle most measuring principles can be automated, or
have already been automated in commercial
respirometers, to measure respiration rate automatically
or even in-line at a wastewater treatment plant. In-line
measurement of respiration rate, however, is out of the
scope of this textbook. One specific method, off-gas
analyses, provides an inherent way to assess the
respiration rate in a pilot-scale or full-scale wastewater
treatment plant. This method is described in Chapter 4.
The information that can be extracted from
respirometric measurements can be divided into two
types: direct and indirect (Spanjers et al., 1998). Direct
information, such as the aerobic respiration rate or
specific methanogenic activity, provides information on
the actual activity of the biomass, and can be used for
example to record respirograms (time series of
respiration rates) in the lab. Indirect information refers to
variables that are deduced from respirometric
measurements, such as microorganism concentration,
substrate concentration and kinetic parameters. In this
case respirometric measurements are used as input to
simple arithmetic calculations or even model fitting.
Chapter 5 will describe how data such as respirometric
data can be used in model fitting to assess deduced
variables and parameters.
Because the objective of this chapter is to provide
only practical guidelines for the assessment of the
respiration rate, only a basic explanation of the
biochemical background will be provided, and the reader
is referred to the literature on biochemistry (Alberts et al.,
2002; Nelson and Cox, 2008). As the practical guidelines
focus on respirometric methods that can easily be carried
out in most laboratories, no rigorous discussion of all the
respirometric measuring principles will be given, and we
© 2016 Henri Spanjers and Peter A. Vanrolleghem. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen, C.M. Lopez-Vazquez and D. Brdjanovic.
ISBN: 9781780404745 (Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
134
Inorganic donors that are converted to their oxidized
form by aerobic microorganisms, where oxygen serves as
the terminal electron acceptor, include ammonium and
nitrite, ferrous (divalent iron) and sulphide, and the
conversions are carried out by nitrifiers (ammonia and
nitrite-oxidizing bacteria), iron-oxidizing bacteria and
sulphide-oxidizing bacteria, respectively. In this case
CO2 forms the source of carbon, and the organisms are
called autotrophs. Non-aerobic microorganisms use
inorganic compounds other than oxygen such as nitrite,
nitrate, sulphate and carbon dioxide as the terminal
electron acceptor. In these cases we are talking about
anoxic (nitrite, nitrate) and anaerobic processes (CO2,
sulphate). Note that in wastewater treatment various
respiration processes may take place simultaneously
where different microorganisms use diverse substrates
and terminal e-acceptors or compete for the same
substrates and terminal electron acceptors.
Product
[CHO]
HCO3–
NH4+
NOX–
H2
e-
Oxidized
During the process of respiration the electron donor
is converted to its oxidized form and the electron
acceptor is converted to its reduced form. In the case of a
carbonaceous donor (organic compounds), the oxidized
form is carbon dioxide. If the electron acceptor is
molecular oxygen then its reduced form is water. The
conversion of a carbonaceous donor with oxygen as the
electron acceptor is carried out by heterotrophic bacteria.
Substrate
H2O
O2
H2O
NOX–
N2
SO42-
S2-
HCO3 –
CH4
Reduced
In biochemical terms, microbial respiration is the
adenosine triphosphate (ATP)-generating metabolic
process in which either organic or inorganic compounds
serve as the electron donor and inorganic compounds
serve as the terminal electron acceptor (e.g. oxygen,
nitrate, sulphate). The universal energy carrier ATP is
generated as electrons removed from the electron donor
are transferred along the electron transport chain from
one metabolic carrier to the next and, eventually, to the
terminal electron acceptor. In this way, microorganisms
convert the energy of intramolecular bonds in the
electron donor to the high-energy phosphate bonds of
ATP (catabolism). The energy is then used to synthesize
the various molecular components required for cell
growth (anabolism), maintenance and reproduction.
Reduced
3.1.1 Basics of respiration
Figure 3.1 shows a schematic overview of some
examples of metabolic conversions. Note that both the
electron donor and the terminal electron acceptor may be
considered as substrate, like many other components that
enter the metabolic pathways. In respirometry,
respiration is generally considered as the consumption of
O2, NO2- or NO3- or (in anaerobic respirometry) the
production of CH4. In general terms, the metabolic
conversions involved in respiration are catabolic
reactions, and several gaseous compounds consumed or
produced during these reactions can be used to assess key
metabolic conversions. Note, however, that in principle
other substances may also be considered, such as the
consumption of NH4+, HS- or S2-, or SO42-, or the
production of N2. Other products that are not included in
the figure but that are also linked to respiration include
H+ and heat and the associated methods are titrimetry and
calorimetry, respectively. However, these are out of the
scope of this chapter.
Oxidized
refer to Spanjers et al. (1998) for a more complete
description of the concepts. For a discussion of the use of
direct and indirect respirometric information in control of
the activated sludge process the reader may consult Copp
et al. (2002).
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Figure 3.1 Schematic overview of some examples of metabolic
conversions. e- denotes the electron that is transferred from the electron
donor to the terminal electron acceptor. [CHO] denotes any carbohydrate.
The coloured substances are generally used as measured variables in
respirometry.
Because the energy generated during the process of
microbial respiration is used for cell growth and
maintenance functions, such as reproduction, cell
mobility, osmotic activity, etc., the respiration rate is
linked to the rate of these processes. However, it is
difficult to differentiate between these two processes. As
an example, consider the aerobic respiration by
heterotrophic microorganisms that use carbonaceous
RESPIROMETRY
(organic) substrate as the electron donor and oxygen as
the terminal electron acceptor. Only a portion (1-Y) of
the consumed organic substrate is oxidised to provide
energy for cell growth and maintenance. The remainder,
typically half (on a weight/weight basis) of the substrate
molecules (the yield Y) is reorganised into new cell mass.
Hence the oxygen consumption rate is linked to the
biomass growth through the yield. In anaerobic
respiration of hydrogenotrophic methanogens, where H2
substrate is used as the electron donor and CO2 as the
electron acceptor, only a small portion of Y of the
substrate is rearranged into biomass, while the largest
part is oxidized to produce CH4.
In activated sludge, carbonaceous substrate removal
is not the only oxygen-consuming process. In addition to
the oxygen consumption by heterotrophic biomass, there
are some other biological processes that may contribute
to the respiration of activated sludge, such as the
oxidation of inorganic compounds by nitrifiers and other
bacteria, and specific microbial oxidation reactions
catalysed by oxidases and mono-oxygenases. Nitrifying
bacteria incorporate only a minor part of the substrate
ammonia into new biomass while most of the substrate
(ammonium) is oxidised for energy production. These
autotrophic bacteria use dissolved carbon dioxide as a
carbon source for new biomass. In comparison to
heterotrophic biomass, nitrifiers need more oxygen for
their growth. Nitrification occurs in two steps: the
oxidation of ammonia to nitrite and the oxidation of
nitrite to nitrate. Like nitrifiers, the autotrophic sulphuroxidizing bacteria and iron-oxidizing bacteria utilise
inorganic compounds instead of organic matter to obtain
energy and use carbon dioxide or carbonate as a carbon
source. Sulphur-oxidizing bacteria are able to oxidise
hydrogen sulphide (or other reduced sulphur compounds)
to sulphuric acid. Iron-oxidizing bacteria oxidise
inorganic ferrous iron to the ferric form to obtain energy.
In addition to bacteria, protozoa and other predating
higher organisms are present in the activated sludge, and
they also consume oxygen. Finally, some inorganic
electron donors such as ferrous iron and sulphide can be
chemically oxidised, also utilising oxygen.
All the above-mentioned oxygen-consuming
processes contribute to the total respiration rate of the
activated sludge. Respirometry is usually intended to
measure only biological oxygen consumption and
sometimes it is attempted to distinguish between
different biological processes such as heterotrophic
oxidation and nitrification. However, in many cases it is
difficult to distinguish between specific microbial
processes and to identify chemical oxygen consumption.
135
O2(g)
CO2(g)
Gas
CH4(g) N (g)
2
CO2(g)
Interface
Liquid
HCO3–
O2
CH4
H2O
N2
NOX-
HCO3– Energy
NOX–
[CHO]
NH4+
H2
Biomass
New biomass
Figure 3.2 The relationship between respiration, substrate utilization and
growth for three types of substrate [CHO], NH4 and H2 and related electron
acceptor e.g. O2, NO3 and HCO3 (Spanjers et al., 1998).
3.1.2 Basics of respirometry
Respirometry is generally defined as the measurement
and interpretation of the rate of biological consumption
of an inorganic electron acceptor under well-defined
experimental conditions. In principle, all the substances
depicted in the ‘substrate-oxidized form’ highlighted in
Figure 3.1 can serve as the measured variable. An
exception is anaerobic respirometry where generally the
production rate of the ultimate reduced product methane
is measured. This is because during anaerobic
degradation many intermediates are involved and it is
impracticable to measure the consumption rate of these
intermediate substrates. Moreover, methanogenesis is
generally not the rate-limiting step; hence, the methane
production rate reflects the rate-limiting process (mostly
hydrolysis in the case of a complex substrate).
Note that in principle one may also measure the
consumption rate of the electron donor, such as [CHO],
NH4+ and H2. However, this is generally not considered
to be respirometry, also because electron donor
substances such as [CHO] and NH4+ may also be
consumed by processes not related to energy generation,
for example the uptake in biomass, and hence not clearly
related to energy generation. Finally, measurement of
CO2 production may be considered respirometry because
CO2 production is related to energy generation (Figure
3.2). However, because CO2 in the gas phase is
associated with the carbonate system, extra
measurements will be needed to measure the pH and
bicarbonate concentration in the liquid phase.
136
Respirometry always involves some technique for
assessing the rate at which the biomass takes up the
electron acceptor (such as O2 and NO3-) from the liquid
or produces its reduced form (such as CH4); see Figure
3.2. For electron acceptors such as O2 and NO3- this is
generally based on measuring the concentration of the
electron acceptor in the liquid phase and solving its mass
balance to derive the respiration rate. If the oxygen
consumption is measured and a gas phase is present, one
has to consider the mass balance of oxygen in the gas
phase as well. Similarly, if measuring the production rate
of methane, the mass balance of methane in both the
liquid phase and the gas phase has to be considered.
For aerobic respirometry i.e. assessing the rate at
which biomass takes up O2, Spanjers et al., (1998)
presented a classification of respirometric principles that
was based on two simple criteria: the first one being the
location of the oxygen measurement, liquid or gas phase;
and the second being the state of the gas and liquid
phases, both either flowing or static. It was found that the
majority of the proposed respirometric devices could be
put into one of the eight classes created by this structure.
Moreover, for each of the classes, examples of
implementations were found in the literature.
3.2 GENERAL METHODOLOGY OF
RESPIROMETRY
3.2.1 Basics of respirometric methodology
The respiration rate is usually measured with a
respirometer. Respirometers range from a simple,
manually operated bottle equipped with a sensor to
complicated instruments that operate fully automatically.
In some cases the bioreactor of the treatment plant itself
can serve as a respirometer. Except for the latter, a feature
common to all respirometers is a reactor, separated from
the bioreactor, where different components (biomass,
substrate, etc.) are brought together. The operation of all
respirometers involves some technique for assessing the
rate at which the biomass takes up a component from the
liquid or produces a component (Figure 3.2). Many
techniques have been developed in the past. However,
Spanjers et al. (1998) found that all measuring techniques
for the respiration rate can be classified into only eight
basic principles according to two criteria: (1) the phase
where the concentration is measured (gas or liquid, G and
L, respectively) and (2) whether or not there is input and
output of liquid and gas (flowing or static, F and S,
respectively). The operation of all existing respirometers
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
can be explained in terms of these criteria. Figure 3.3
shows a generic scheme for a respirometer. Note that the
gas phase also includes bubbles dispersed in the liquid
phase. In the subsequent sections, the principles will be
discussed according to the above criteria. We will not
discuss the usefulness of the different measuring
techniques, because we believe that any technique has its
merits, depending on the specific application, provided
that the correct measuring conditions are satisfied.
Gas
Liquid
Figure 3.3 Generic scheme of a respirometer.
3.2.2 Generalized principles: beyond oxygen
3.2.2.1 Principles based on measuring in the liquid
phase
The majority of the techniques based on measurement in
the liquid phase use a specific electrode or sensor. A
reliable respiration rate measurement is only possible if
the sensor is correctly calibrated and if a number of
environmental variables, such as temperature and
pressure, is accounted for. Sensors also have a response
time that must be accounted for in some respirometric
setups.
Respirometers that are based on measuring dissolved
oxygen (DO) concentration in the liquid phase use a DO
mass balance over the liquid phase. Consider a system
consisting of a liquid phase, containing biomass, and a
gas phase both being ideally mixed and having an input
and output (Figure 3.4). It is assumed that the DO
concentration in the liquid phase can be measured. The
DO mass balance over the liquid phase is:
d ( VL· SO2 )
= Qin · SO2,in − Qout · SO2 +
dt
VL· kLa ⋅ ( S*O2 − SO2 ) − VL · rO2
Eq. 3.1
Where, SO2 is the DO concentration in the liquid
phase (mg L-1), S*O2 is the saturation DO concentration in
the liquid phase (mg L-1), SO2,in is the DO concentration
RESPIROMETRY
137
in the liquid phase entering the system (mg L-1), kLa is
the oxygen mass transfer coefficient, based on liquid
volume (h-1), Qin is the flow rate of the liquid entering the
system (L h-1), Qout is the flow rate of the liquid leaving
the system (L h-1), rO2 is the respiration rate of the
biomass in the liquid, (mg L-1 h-1), and VL is the volume
of the liquid phase (L).
DO
ΔSO2/Δt = ‒rO2. Typical of this principle is that the DO
may become exhausted after some time so that for
continued measurement of rO2 reaeration it is necessary
to bring the DO concentration back to a higher level. DO
and substrate limit the respiration when their
concentrations become too low, causing a non-linear DO
decrease complicating the assessment of the differential
term. Note that in Figure 3.5 there is a gas phase.
However, it is assumed there is no mass transfer from the
gas phase into the liquid phase. In practice, in order to
prevent the input of oxygen into the liquid, the gas phase
may be absent. The procedure for the determination of rO
according to standard methods (APHA et al., 2012) is
based on this principle.
DO
DO
Figure 3.4 The liquid phase principle, flowing gas, flowing liquid (LFF).
Notice that, since it is a mass balance over the liquid
phase, Eq. 3.1 does not contain gas flow terms. The first
and second term on the right-hand side represent the
advective flow of DO in the input and output liquid
streams. In most systems Qin and Qout will be equal so that
the liquid volume is constant. The third term describes
the mass transfer of oxygen from the gas phase to the
liquid phase. The last term contains the respiration rate to
be derived from the mass balance. Therefore, SO must be
measured and all other coefficients be known or
neglected (i.e. non-influential). In practice, the
determination of rO2 can be simplified in several ways. In
what follows it is assumed that the liquid volume is
constant, so that the terms in Eq. 3.1 can be divided by
VL.
• Static gas, static liquid (LSS)
One approach is to use a method without liquid flow and
oxygen mass transfer (Figure 3.5). Then the first three
terms on the right-hand side of Eq. 3.1 fall away and the
mass balance reduces to:
dSO2
= − rO2
dt
Eq. 3.2
Hence, to obtain the respiration rate only the
differential term has to be determined. This can be done
by measuring the decrease in DO as a function of time
due to respiration, which is equivalent to approximating
the differential term with a finite difference term:
Figure 3.5 The liquid phase principle, static gas, static liquid (LSS).
• Flowing gas, static liquid (LFS)
The disadvantage of the need for reaerations can be
eliminated by continuously aerating the biomass. Then,
the oxygen mass transfer term kLa ⋅ (S*O2 − SO2 ) must be
included in the mass balance (Eq. 3.3):
dSO2
= kLa ⋅ (S*O2 − SO2 ) − rO2
dt
Eq. 3.3
To obtain rO2, both the differential term and the mass
transfer term must be determined. To calculate the latter,
the mass transfer coefficient (kLa) and the DO saturation
concentration ( S*O2 ) must be known. These coefficients
have to be determined regularly because they depend on
environmental conditions such as temperature,
barometric pressure and the properties of the liquid
(viscosity, salinity, etc.). The simplest approach is to
determine these by using separate reaeration tests and
look-up tables. Another approach is to estimate the
coefficients from the dynamics of the DO concentration
response by applying parameter estimation techniques.
The advantage of the latter method is that the values of
138
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
the aeration coefficients can be updated relatively easily.
This respirometric principle allows the measurement of
rO2 at a nearly constant DO concentration, thereby
eliminating the dependency of rO2 on the DO
concentration (provided DO >> 0 mg L-1). Note that,
whereas Figure 3.6 shows an input and an output in the
gas phase, there is no gas flow term in Eq. 3.3. There is
no need to consider gas flow terms provided S*O2 is
known or determined.
and VL are instrument constants and are, therefore,
assumed to be known or calibrated. This principle is in
fact the continuous counterpart of the one explained in
Eq. 3.2, and it is as such also sensitive to the effect of
substrate and DO limitation. However, the effect of
limiting substrate can be eliminated by the continuous
supply of substrate (wastewater) and DO to the
respiration cell.
• Flowing gas, flowing liquid (LFF)
Without the above simplifications the full mass balance
(Eq. 3.1) holds for the principle depicted in Figure 3.4.
To obtain respiration rate measurements with this
principle, a combination of the approaches mentioned for
the above simplified principles is required. For instance,
the flow rates and the inlet oxygen concentrations must
be measured, while the coefficients kLa and S*O2 must be
DO
assessed, e.g. by estimating these from the dynamics of
the DO concentration.
Figure 3.6 The liquid phase principle, flowing gas, static liquid (LFS).
• Static gas, flowing liquid (LSF)
Repetitive aeration or estimation of oxygen transfer
coefficients, as with the above principles, can be avoided
when liquid with a high enough input DO concentration
flows continuously through a closed completely mixed
cell without the gas phase (Figure 3.7). The liquid flow
terms now have to be included in the mass balance (Eq.
3.4):
dSO2 Qin
Q
=
· SO2,in − out · SO2 − rO2
dt
VL
VL
Eq. 3.4
DO
DO
Figure 3.7 The liquid phase principle, static (no) gas, flowing liquid (LSF).
Both DO concentrations, SO2,in and SO2, must be
measured to allow calculation of rO2. In a respirometer Qin
3.2.2.2 Principles based on measuring in the gas
phase
Respirometric techniques based on measuring gaseous
oxygen always deal with two phases: a liquid phase
containing the respiring biomass and a gas phase where
the oxygen measurement takes place. The main reason
for measuring in the gas phase is to overcome difficulties
associated with interfering contaminants common in the
liquid phase (e.g. the formation of biomass film on the
sensor). Gaseous oxygen is measured by physical
methods such as the paramagnetic method, or gasometric
methods.
Gasometric methods measure changes in the
concentration of gaseous oxygen. According to the ideal
gas law P·V = n·R·T, these can be derived from changes
in the pressure (if volume is kept constant, the
manometric method) or changes in the volume (if
pressure is kept constant, the volumetric method). These
methods are typically applied to closed measuring
systems (no input and output streams), which may
provoke a need for reaerations and thus temporary
interruption of the measurements. This limits the
possibility for continued monitoring of the respiration
rate. However, interruptions because of reaerations are
not needed if the consumed oxygen is replenished at a
known rate, e.g. by supplying pure oxygen from a
reservoir or by using electrolysis. The rate at which
oxygen is supplied is then equivalent to the biological
respiration rate (assuming infinitely fast mass transfer to
the liquid). Because carbon dioxide is released from the
RESPIROMETRY
139
liquid phase as a result of the biological activity, this gas
has to be removed from the gas phase in order to avoid
interference with the oxygen measurement. In practice
this is done by using alkali to chemically absorb the
carbon dioxide produced.
S*O2 = H ⋅ CO2
Respirometric principles based on measuring gaseous
oxygen also use oxygen mass balances to derive the
respiration rate. However, in addition to the mass balance
in the liquid phase (Eq. 3.1), a balance in the (ideally
mixed) gas phase must be considered (Figure 3.8):
and that:
d
( VG · CO2 ) = Fin · CO2,in − Fout · CO2 − VL· kLa ⋅ (S*O2 − SO2 )
dt
….Eq. 3.5
Where, CO2 is the O2 concentration in the gas phase
(mg L-1), CO2,in is the O2 concentration in the gas entering
the system (mg L-1), Fin is the flow rate of the gas entering
the system (L h-1), Fout is the flow rate of the gas leaving
the system (L h-1), and VG is the volume of the gas phase
(L).
O2
O2
Eq. 3.6
it is reasonable to state that:
SO2 = H ⋅ CO2
Eq. 3.7
dSO2
dC
= H ⋅ O2
dt
dt
Eq. 3.8
Hence, the measurement in the gas phase is a good
representation of the condition in the liquid phase,
provided the proportionality (Henry) constant H is
known, e.g. from calibration or tables, and the mass
transfer coefficient is high. The validity of this
equilibrium assumption should be critically evaluated.
• Static gas, static liquid (GSS)
The simplest gas phase technique for measuring the
respiration rate is based on a static liquid phase and a
static gas phase, i.e. no input or output (Figure 3.9). In
addition to the DO mass balance in the liquid phase, an
oxygen mass balance in the gas phase must be
considered:
dSO2
= kLa ⋅ (S*O2 − SO2) − rO2
dt
Eq. 3.9
d(V G· CO2)
= - V L· kLa ⋅ (S*O2 − SO2)
dt
Figure 3.8 The gas phase principle, flowing gas, flowing liquid (GFF).
The term VL·kLa·( S*O2 − SO2 ) represents the mass
transfer rate of oxygen from the gas phase to the liquid
phase, and it is the connection between the two phases.
From mass balances (Eq. 3.1 and Eq. 3.5) it follows that,
in order to calculate rO2, CO2 must be measured (directly
or using the gas law, see above) and knowledge of SO2 is
required. However, SO2 is not measured in the gas phase
principles. In these respirometric principles it is assumed
that the oxygen concentrations in the gas and liquid
phases are in equilibrium, i.e. mass transfer is sufficiently
∗
fast (kLa → ∞), so that SO2 ≈ SO2
. Since, by definition,
the saturation DO concentration is proportional to the O2
concentration in the gas phase:
Eq. 3.10
Hence, in order to calculate rO2, the change of the
oxygen concentration in the gas phase, dCO2/dt, must be
measured and knowledge of dSO2/dt is required (Eq. 3.9).
It is possible to measure dCO2/dt by using an oxygen
sensor. If a gasometric method is used, dCO2/dt is related
to the change in volume or the change in pressure (Eq.
3.10.
O2
Figure 3.9 The gas phase principle, static gas, static liquid (GSS).
140
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
With this principle, the same restriction as with the
simplest DO-based principle exists: when the oxygen
becomes exhausted it must be replenished by, for
instance, venting the gas phase in order to continue the
measurement of rO2.
• Static gas, flowing liquid (GSF)
Implementations of the gas phase principle with static gas
and flowing liquid (Figure 3.11) have not been found in
literature or in practice so far.
• Flowing gas, static liquid (GFS)
Another technique is based on a flowing gas phase, i.e.
the biomass is continuously aerated with air (or pure
oxygen) so that the presence of sufficient oxygen is
assured (Figure 3.10). In comparison to Eq. 3.10, two
transport terms must be included in the mass balance on
the gas phase:
dSO2
= kLa ⋅ (S*O2 − SO2) − rO2
dt
3.11
O2
Eq.
Figure 3.11 The gas phase principle, static gas, flowing liquid (GSF).
d(V G· CO2)
= Fin· CO2,in − Fout· CO2 − V L ⋅ kLa ⋅ (S*O2 − SO2)
dt
Eq. 3.12
Figure 3.10 The gas phase principle, flowing gas, static liquid (GFS).
• Flowing gas, flowing liquid (GFF)
The gas phase principle can also be applied to a full-scale
bioreactor. In this case there are liquid input and output
streams for the reactor, and transport terms must be added
to the mass balance in the liquid phase (Eq. 3.1). The
assumption on proportionality between CO2 and SO2 (Eq.
3.7) becomes more critical because, in addition, the
liquid outflow term also depends on it. Additional
measurement of dissolved oxygen may then be useful for
a correct assessment of the respiration rate. The
technique then would no longer be a pure gas phase
principle. Note, however, that in general combining L
and G principles may lead to more reliable respiration
rate measurements.
In order to allow the calculation of rO, the gas flow
rates, Fin and Fout, and the oxygen concentrations in the
input and output streams, CO,in and CO, must be known in
addition to the variables of the previous technique. Of
these, usually CO is measured and the others are set or
known. A gasometric method is not evident here, and the
measurement of CO is done for example with the
paramagnetic method.
Table 3.1 summarises the eight measuring principles.
The first column contains the names of the mass balance
terms, and the second column the mathematical
equivalents. The succeeding columns list the
respirometric principles, the first four being the liquidphase principles, and the others being the gas-phase
principles. The mass balances for each principle are
formed by multiplying the mathematical terms with the
coefficients in the column of the appropriate principle
and summing them up.
O2
RESPIROMETRY
141
Table 3.1 Overview of measuring principles of respiration rates.
Respirometric principle →
Process ↓
Equation ↓
Respiration
VL· rO2
Dissolved oxygen accumulation
d
(VL · SO 2 )
dt
Qin · SO2,in − Qout ·SO2
Figure nr. →
Liquid flow
*
O2
− SO 2 )
Gas exchange
VL · k L a (S
Gaseous oxygen accumulation
d
(VG · C O 2 )
dt
Fin · CO2,in − Fout · CO2
Gas flow
Gas exchange
*
O2
VL · k L a (S
Measurement in LIQUID phase
LSS
LFS
LSF
LFF
3.5
3.6
3.7
3.4
-1
-1
-1
-1
Measurement in GAS phase
GSS
GFS
GSF
GFF
3.9
3.10
3.11
3.8
-1
-1
-1
-1
-1
-1
-1
-1
-1
1
1
1
1
3.3.1 Equipment for anaerobic respirometry
-1
1
1
1
1
1
1
-1
-1
-1
-1
-1
-1
1
-1
-1
• Measuring the biogas composition and correcting the
measured flow
Measuring the biogas composition can be done with gas
chromatography. However, there are also cheaper and
easier methods, for example, the apparatus shown in
Figure 3.12.
9
8
7
6
5
4
3
2
1
9
8
7
6
5
4
3
2
1
9
8
7
6
5
4
3
2
1
CH4
Inject
biogas
CH4 + CO2
To carry out an anaerobic respirometric test there are two
requirements. Firstly, a setup is needed in which
anaerobic respiration takes place. This can be a small
bottle or larger reactor. In this bottle or reactor a
substrate, for example, primary sludge or starch, and an
inoculum with the consortia required for anaerobic
respiration are combined. Secondly, a system to measure
the methane production is required. To quantify
anaerobic respiration, the flow of electrons has to be
determined. Neither the consumption of substrate nor the
consumption of an electron acceptor can be measured
directly and therefore the final products of anaerobic
respiration, H2 and methane, are determined in anaerobic
respirometry.
-1
1
− SO 2 )
3.3 EQUIPMENT
-1
3.3.1.1 Biogas composition
Methane leaves the bottle or reactor via the biogas.
Besides methane, biogas also contains CO2, H2S and
traces of other compounds. Thus, in order to quantify the
methane flow, both the biogas flow and the biogas
composition need to be known. To this end, the
composition of biogas can be either measured or adapted.
Adaptation of the biogas means removing all gas other
than methane prior to flow quantification.
Figure 3.12 An inexpensive and simple way for determining the methane
concentration in biogas.
This tube is filled with an alkaline solution (typically
3 molar of NaOH) to remove CO2 and H2S, which
dissolve in an alkaline solution, leaving only methane in
the gas phase. First, biogas is injected into the left leg (t
= 0). Then CO2 and H2S dissolve in the alkaline solution
(orange) over time until all the CO2 and H2S are removed
142
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
(t = end). The methane content can be calculated from the
differences in volume at t = 0 and t = end. In this example
the total biogas volume is 10 mL, of which 4 mL is
dissolved, hence, the methane content of the biogas
sample is calculated as 60 %. The time required for all
the CO2 and H2S to be absorbed into the alkaline solution
is to be determined experimentally. This can be done by
assessing the time required for the system to reach steady
state, i.e. the gas volume does not change anymore. To
calculate the CH4 flow rate from the bottle or reactor, the
volume should be corrected using the ideal gas law and
the actual temperature.
• Removing other gases from the biogas
When CO2 and H2S are removed, the gas will usually
contain 100% methane (some N2 and H2 may be present).
This means that in situ removal of compounds other than
methane, combined with flow measurement, yields the
methane flow. In practice, this means that from the
reaction vessel, the gas is led over a large surface of an
alkaline solution (typically 3 molar NaOH solution)
(Figure 3.13). Notice that, in contrast to what is shown
in Figure 3.13, the inlet to the scrubber bottle may not be
submerged. This is to prevent back flow of alkaline
solution. For example, when there is an under pressure in
the head space of the reaction vessel, alkaline solution
would be sucked into this reaction vessel, compromising
the experiment instantly. An under pressure can occur
when the temperature of the head space drops, e.g. when
a thermostatic water bath fails or the door of the incubator
is left open for a while.
Reaction
vessel
To gas
flow meter
3M NaOH
Figure 3.13 Schematic picture of the scrubber bottle.
3.3.1.2 Measuring the gas flow
There is a large variety of ways to measure gas
production but in lab-scale anaerobic respirometry it is
usually limited to manometric or volumetric methods.
• Manometric methods
Manometric methods are based upon measuring pressure
increase in the head space of a reaction vessel. As biogas
is produced, the pressure in the headspace increases.
However, high headspace pressure can result in increased
CO2 solubility, which may significantly disturb microbial
activity (Theodorou et al., 1994). Therefore, the pressure
needs to be released periodically to prevent it from
becoming too high (pressure release). Generally, an
upper limit of 1.4 bar is applied. One must also make sure
that the reaction vessel is designed to withstand the
pressure. When the pressure is not released
automatically, this method requires labour during the
experiment. Inappropriate operation can lead to
explosion of the reaction vessel. It is strongly
recommended to always use safety goggles when
working with the manometric method. In addition, the
initial pressure measurement and the measurements after
draining should also consider the temperature effect on
gas pressure and water vapour pressure, i.e. an
equilibrium condition shall be reached in the reaction
vessels before measurement. To calculate the gas
production, the pressure increases between two pressure
drains are summed and with the ideal gas law the amount
of produced moles of biogas is calculated from this total
increase in pressure. The composition of the biogas needs
to be measured as well if there is no CO2 and H2S
scrubbing.
• Volumetric methods
A classical and robust volumetric method is to use
Mariotte’s bottle (McCarthy, 1934), where the gas is
introduced into a bottle with an outlet for the liquid and
it displaces the liquid (Figure 3.14). The weight or
volume of the displaced liquid indicates the volume of
gas that is produced. When the liquid is an alkaline
solution, CO2 or H2S is scrubbed in situ and the weight
of displaced liquid will indicate the volume of methane.
A disadvantage is that the flask needs to be refilled
periodically, which disrupts the pressure control. If the
displaced liquid is measured with a balance connected to
a computer, the gas production can be measured in real
time.
RESPIROMETRY
143
3.3.2 Equipment for aerobic and anoxic
respirometry
h
Figure 3.14 Mariotte’s bottle for measuring produced gas volumes.
Another example of a volumetric measuring principle
is the tilting mechanism (Figure 3.15).
Gas output
Counter
Gas input
Liquor chamber
Liquor casing
Gas
Tilting box
(measuring chamber)
Output nozzle
Figure 3.15 Schematic overview of the cross section of a tilting box
anaerobic respirometer. The tilting box alternates between right and left
as the gas is introduced. The amount of clicks is measured and registered
(www.ritter.de).
The advantage is that there is no need for actively
resetting the gas flow meter. It can run continuously
without requiring attention as Mariotte’s bottle does.
Several commercial systems exist that use this principle.
The measurement principle is based on a tilting box
submerged in oil or water. This box is filled from the
bottom with gas. This gas accumulates under the
chamber and at a certain point this gas induces positive
buoyancy and then the box tilts, releasing the gas and
thus resetting the system. Every tilt is counted and from
this a gas volume is calculated. The downside of this
method is that it is rather expensive and that it has a
limited flow range (up to 4 L h-1).
Similar to anaerobic respirometry, a setup is needed in
which aerobic and anoxic respiration takes place. This
basically consists of a stirred bottle or reactor where
biomass under aerobic or anoxic conditions and
wastewater, or a specific substrate, are combined. In
addition, an arrangement is required to measure the
uptake of the terminal electron acceptor, i.e. oxygen,
nitrite or nitrate. Data handling may be manually, as is
mostly the case in BOD measurements, or completely
automated, for example if the measured data is to be
converted to respiration rates with a high measuring
frequency. In a number of sophisticated (including
commercial) respirometers the operation of the
equipment is so complicated that it requires an automated
control system.
3.3.2.1 Reactor
The reactor is usually a vessel with a volume ranging
from a few 100 mL to several litres. Depending on the
application, laboratory or field, the material may be glass
or plastic, and is often transparent in order to enable
inspection of the content. Depending on the measuring
principle (Section 3.3.1.2), the vessel is completely
sealed to prevent the exchange of oxygen with the gas
phase, or open to allow the transfer of oxygen from the
gas phase. Open vessels may also be equipped with
aeration equipment (e.g. a sparger) to enhance oxygen
transfer. In some cases the vessel may be operated in both
open mode (for aeration) and closed mode (for measuring
oxygen uptake). In all cases, the vessel is completely
mixed, with a magnetic bar, an impeller, a pump, or by
aeration. In the laboratory the vessel may be
thermostated, often using a double wall for
cooling/heating or just a heating element if the
temperature is maintained above ambient temperature.
Depending on the operation principle (flowing liquid,
flowing gas), the reactor may have several inlet and outlet
ports, and one or more apertures to accommodate (a)
sensor(s). Supplementary equipment may include valves,
pumps (for biomass, wastewater, substrate, air, and gas),
a mixing tank, substrate container, oxygen container,
NO3 supply container, sample pre-treatment unit (sieve,
filter), oxygen generator, etc.
3.3.2.2 Measuring arrangement
In many cases the measuring arrangement consists of a
sensor (i.e. a probe with an associated meter whether or
144
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
not connected to a data acquisition system) to measure
the concentration of the electron acceptor, i.e. oxygen,
nitrite or nitrate. Oxygen may be measured directly in the
liquid phase with a galvanic, polarographic or optical
dissolved oxygen probe. In simple laboratory tests,
especially a BOD test, dissolved oxygen may be
measured with a titrimetric or photometric method
(Section 3.4.2). The oxygen concentration in the gas
phase may be directly measured with a paramagnetic
oxygen analyser. However, changes in oxygen
concentration may be measured by means of a pressure
sensor or gas volume displacement sensor. Nitrate and
nitrite concentrations in the liquid phase can be measured
with an ion-selective or UV-spectrophotometric sensor
(Rieger et al., 2008).
operation of all the existing respirometers can be
explained in terms of this classification and the
corresponding mass balances. However, only a limited
number of respirometers have been applied in
considerable numbers in research and practice, or even
commercial production. In what follows we describe
some respirometers in terms of their basic principle and
technical implementation. However, it should be
emphasized that by no means should this be understood
as a recommendation for a specific method. The choice
of a certain measuring principle, its technical
implementation, or commercial manifestation depend on
the measurement purpose, skill of the user and available
budget.
Sensors may have slow response times, and it is
important to ensure that the sensor is fast enough to
follow the kinetics of the biochemical process. As a rule,
the sensor must be 10 times faster than the measured
reaction rate.
The LSS principle can be considered as the simplest
respirometric principle because the absence of flowing
liquid and gas implies that no supplementary materials,
such as pumps or aeration equipment, are needed. The
BOD test (Section 3.4.2) is an example of the application
of this principle. Figure 3.16 shows an example of a BOD
bottle used in a test where DO is only measured in the
beginning and at the end of the test, and an example of a
BOD bottle with continuous measurement of the oxygen
uptake by means of a pressure sensor. The latter allows
for the assessment of the ultimate BOD and the first order
oxygen uptake rate coefficient (Section 3.4.2).
3.3.2.3 Practical implementation
Many practical implementations have been described in
the literature and a number of them have been introduced
onto the market. As explained in Section 3.3.1, all the
measuring techniques for the respiration rate can be
classified into only eight basic principles and the
• Liquid phase, static gas, static liquid (LSS) principle
Figure 3.16 A BOD bottle for the classical BOD test (left) and bottles for the continuous measurement of oxygen uptake (right) (photos: Wheaton and VELP
Scientifica).
RESPIROMETRY
145
However, the LSS principle has also been
implemented in a semi-continuous version to measure the
respiration rate of biomass semi-continuously, both in the
lab and in the field. Because of the much higher biomass
concentration than used in a typical BOD test, the DO
concentration drops due to respiration within a few
minutes from a near saturated concentration to a limiting
concentration. The respiration rate is then calculated
from the slope of the DO concentration decline. To allow
repeated measurement of the respiration rate the biomass
is reaerated after each measurement, which yields a
typical saw-tooth DO profile (Figure 3.17). In this
example the time between the on/off periods of aeration
is constant. Other respirometers make the aeration switch
on and off dependent on the actual DO concentration, e.g.
between 4 and 6 mg O2 L-1. These upper and lower DO
limits should be defined carefully; they determine the
frequency of respiration rate date and their accuracy.
Indeed, when the respiration rate is low, it may take a
long time to lower the DO concentration from the upper
to the lower limit, whereas too short declines make the
calculation of the respiration rate sensitive to
measurement errors since only a few DO data points are
available.
reaeration phase to the DO decline phase can take some
time (tens of seconds) and is affected by the removal of
gas bubbles from the liquid and by the transient response
of the DO probe. Respirometers based on this technique
allow the measurement of the respiration rate with a
measuring interval ranging from typically a few minutes
to several tens of minutes. They also permit, especially in
the lab, the generation of respirograms by the addition of
wastewater or specific substrates. Figure 3.18 is an
illustration of a respirometer based on this technique.
This floating ball respirometer is designed for automated
sampling and discharge of activated sludge, repeated
aeration and calculation of respiration rate.
8.0
7.5
7.0
O2 (mg L-1)
6.5
A)
6.0
5.5
5.0
B)
4.5
4.0
3.5
3.0
21.5
22.0
22.5
23.0
23.5
24.0
24.5
25.0
25.5
Time (h)
Figure 3.17 Raw DO concentration data from a LSS respirometer with reaeration.
Potential difficulties with this technique are that
during the DO decline, oxygen transfer from the gas
phase to the liquid phase needs to be avoided (especially
critical when the respiration rate is low) and that
identifying a linear decrease of DO is not always evident.
The latter is especially a challenge when the measuring
technique is automated. In fact, the transition from the
Figure 3.18 Example of a practical implementation of respirometry
following the LSS principle. (A) a close up with a DO probe visible and (B)
the respirometer in place, i.e. floating on the activated sludge in an
aeration tank (photos: Strathkelvin Instruments Ltd.).
146
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
interrupting the aeration, adding hydrogen peroxide, or
even by the addition of a readily biodegradable substrate.
The obtained reaeration curve can then be used to
estimate the kLa and DO saturation concentration S*O2 .
Nonlinear parameter estimation techniques as presented
in Chapter 5 are recommended to obtain reliable values.
The advantage of the method based on disturbance by
respiration of a readily biodegradable substrate is that the
values of the aeration coefficients can be updated
relatively easily and frequently. However, when only a
moderate DO disturbance occurs, the accuracy of the
estimated kLa is low. Also, it must be assumed that the
respiration rate has dropped to a constant (endogenous)
rate during the reaeration part of the curve.
Figure 3.19 Example of another practical implementation following the
LSS principle (photo: P.A. Vanrolleghem).
Figure 3.19 is an example of another practical
implementation following the LSS principle with a
closed respiration cell (the right-hand vessel) that is filled
with activated sludge from the aerated tank (the left-hand
vessel). The DO decline in this cell is measured until a
certain minimal DO is reached (or after a given time, or
a given DO variation), after which the content is
exchanged for fresh, aerated activated sludge from the
aerated vessel and a new cycle is started.
• Liquid phase, flowing gas, static liquid (LFS) principle
The disadvantage of the need for reaeration can be
eliminated by continuously aerating the biomass. The
continuous supply of oxygen guarantees a non-limiting
DO concentration even at high respiration rates, for
example at high biomass concentration and high
wastewater or substrate doses. Moreover, continuous
aeration allows an open vessel, which facilitates the
addition of wastewater and substrate.
To obtain the respiration rate, both the mass transfer
term (under process conditions) and differential term in
the DO mass balance must be known (Eq. 3.3). The mass
transfer term is calculated from the measured DO
concentration, the mass transfer coefficient kLa and the
DO saturation concentration S*O2 . These two coefficients
have to be determined regularly because they depend on
environmental conditions such as temperature,
barometric pressure and the properties of the liquid (e.g.
salts and certain organics). The simplest approach is to
determine these by using separate reaeration tests and
look-up tables. Standard procedures for these tests under
process conditions are available and are based on the
disturbance of the equilibrium DO concentration by
Estimation of S*O2 is not required when one is only
interested in the substrate-induced respiration, i.e.
exogenous respiration rO2,exo. Total respiration is the sum
of endogenous respiration rO2,endo and exogenous
respiration rO2,exo. Considering rO2,endo, kLa and S*O2 to be
constant over a short interval, it can be shown that the
equilibrium DO concentration reached under endogenous
conditions S*O2,endo encapsulates the endogenous
respiration (Kong et al., 1996). The mass balance for
oxygen can then be rewritten as:
dSO2
= kLa · S*O2 ‒ SO2 ‒ rO2 =
dt
kLa · S*O2 ‒ SO2 ‒ rO2,endo ‒ rO2,exo
Eq. 3.13
By putting rO2,endo = kLa (S*O2,endo ‒ SO2) and replacing
rO2,endo in the above equation, one obtains:
dSO2
= kLa · S*O2,endo ‒ SO2 ‒ rO2,exo
dt
Eq. 3.14
To estimate the exogenous respiration rate rO2,exo from
this mass balance, one therefore only needs to estimate
the equilibrium DO concentration S*O2,endo (directly from
the data, Figure 3.20) and the kLa from the reaeration part
of a DO disturbance curve obtained with a readily
biodegradable substrate. Another advantage of this LFS
respirometric principle is that it allows the measurement
of rO2 at a nearly constant DO concentration, thereby
eliminating the dependency of the respiration on the DO
concentration (provided DO >> 0 mg L-1). Yet another
advantage is that the time interval at which respiration
rate values can be obtained is short, and is in fact only
limited by the measuring frequency of the DO probe.
This makes a respirometer based on the LFS principle
suitable for kinetic tests and model optimization
experiments.
RESPIROMETRY
147
A)
Interface
Air
Discharge
Calibration
addition
Wastewater
addition
Drain
B)
2
r O2,exo (mg O L-1 min-1)
1.50
1.25
C+N
1.00
C
0.50
N
0.25
0.00
0
30
60
90
120
Time (min)
Figure 3.20 (A) Diagram showing a LFS principle-based respirometer
setup (Vanrolleghem et al., 1994) and (B) an example of typical raw data
(Kong et al., 1996).
Obviously, whereas in its basic form a respirometer
following the LFS principle consists of a vessel equipped
with an aerator, an impellor or recirculation pump for
mixing and a DO probe, a more advanced version
designed for automated (online) experiments includes
supplementary equipment such as pumps for filling the
vessel with biomass, wastewater and substrate addition,
level switches, and valves for vessel drainage. This
requires a suitable data handling and control system,
besides sufficient computing capacity for the estimation
procedure. Figure 3.21 shows an example of a
commercial version of this measuring principle.
Figure 3.21 Example of a commercial version of a LFS principle-based
respirometer using the setup depicted in Figure 3.20. In the middle is the
thermostated vessel and at the bottom, left and right, pumps for
wastewater and calibration addition (photo: Kelma NV).
• Liquid phase, static gas, flowing liquid (LSF) principle
The LSF respirometric measuring principles allow
continuous sampling of a biomass stream, for example
activated sludge from a lab reactor or full-scale aeration
tank, while measuring the respiration rate. When the
biomass flows through a closed completely mixed vessel
(respirometric cell) without a gas phase, then, following
Eq. 3.4, the respiration rate can be calculated if the
volume of the respirometric cell and the flow through rate
are known and DO concentrations at the inlet and outlet
of the cell are measured. Alternatively, the DO can be
measured in the source reactor (provided that the
148
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
decrease in DO in the supply line is negligible) and in the
respirometric cell itself. In any case, as with other
principles the DO concentration in the cell must be high
enough to prevent DO limitation, which requires a
sufficiently high input DO concentration at a given
respiration rate. A potential source of erroneous
measurements is associated with the use of two DO
probes to measure DO at the inlet and outlet, because
when probe characteristics differ slightly, relatively large
relative errors can occur in the difference between the
two DO measurements needed to calculate the respiration
rate. Spanjers and Olsson (1992) solved this by
measuring the DO concentrations in the inlet and outlet
of the cell alternately with one single probe located at one
port. This was realized by periodically changing the flow
direction through the vessel using four solenoid valves
which were activated two by two (Figure 3.22).
If a steady state is assumed with respect to the
respiration rate then the rate can be calculated using Eq.
3.4 by assuming that the derivative is zero. However, by
approximating the equation by a difference equation, the
respiration rate can be calculated under dynamic
conditions.
Figure 3.23 shows a commercial version of the LSF
principle using the one probe solution. This respirometer
measures the respiration rate with an interval of typically
one minute and can be connected to a lab reactor or a fullscale aeration tank, for example by using a fast loop.
DO probe
A)
DO
DO
Stirrer
In
Valves
Out
B)
10
Figure 3.23 Example of a commercial version of the LSF principle using the
setup depicted in Figure 3.22. On the left is a lab-scale aeration tank with
activated sludge. In the equipment box on the left is the sampling pump,
and on the right is the set of solenoid valves. The vessel is placed behind
the valves (photo: Applitek NV).
DO (mg O2 L-1)
Sample
8
DO in
6
DOout
4
2
0
0
0.00
0.25
0.50
0.75
1.00
Time (h)
Figure 3.22 Example of a practical implementation following the LSF
principle, with one DO probe to measure DO concentration at the inlet and
outlet of the respiration vessel (Spanjers, 1993). (A) a schematic of the
measurement arrangement and (B) a typical profile of a signal recorded by
the single probe (Spanjers and Olsson, 1992). The signal represents the DO
oscillating between the DO at the inlet and the DO at the outlet of the cell.
This DO signal is the basis for the calculation of the respiration rate.
• The gas phase, static gas, static liquid (GSS) principle
Like the LSS principle which is one of the liquid phase
principles, the GSS principle is the simplest gas phase
principle because the absence of flowing gas and liquid
implies that no supplementary materials, such as pumps
and aeration equipment, are needed. However, because
the calculation of the respiration rate is based on the
measurement of oxygen in the gas phase and the actual
respiration takes place in the liquid phase, the relation
between the gas phase oxygen dynamics must be related
to the respiration rate. Thus, in addition to the DO mass
balance in the liquid phase, an oxygen mass balance in
the gas phase must be considered and the set of equations
solved for the respiration rate, assuming a transfer
relation between the gas and liquid.
RESPIROMETRY
A typical GSS-based respirometer consists of a
biomass vessel, mixing equipment and a gas
measurement arrangement. Gaseous oxygen is measured
by physical methods, such as the gasometric method or
the paramagnetic method. Gasometric methods measure
changes in the concentration of gaseous oxygen, which
can be derived from changes in the pressure (if volume is
kept constant, a manometric method) or changes in the
volume (if pressure is kept constant, a volumetric
method, e.g. Mariotte’s bottle), see Section 3.3.1.2. As
with the LSS principle, when the oxygen consumption is
too high, these methods need replenishment of the
gaseous oxygen and thus temporary interruption of the
measurements. This limits the possibility for continued
monitoring of the respiration rate. An important
complication of the GSS principle is that, because carbon
dioxide is released from the liquid phase as a result of the
biological activity, this gas has to be removed from the
gas phase in order to avoid interference with some of the
simpler oxygen measurement principles. In practice this
is done by using alkali to chemically absorb the carbon
dioxide produced.
Similarly to the LSS principle, the GSS principle can
be used to carry out a BOD test (Section 3.4.2), for
example Oxitop (Figure 3.16).
• The gas phase, flowing gas, static liquid (GFS) principle
Like the LFS measuring principle, respirometers using
the GFS principle are based on a flowing gas phase, i.e.
the biomass is continuously aerated with air (or pure
oxygen) so that the presence of sufficient oxygen in the
liquid is ensured. However, the calculation of the
respiration rate is based on measurement of oxygen in the
gas phase, more specifically the gas leaving the liquid
phase after aeration, also called off-gas. Because the offgas may contain other components that are influenced by
the metabolic processes in the liquid phase, such as
carbon dioxide and nitrogen, it is obvious that these may
also be measured to obtain additional information on the
activity of the biomass.
A typical GFS-based respirometer consists of a
biomass vessel, aeration and mixing equipment and an
off-gas measurement arrangement. Gaseous oxygen is
measured by physical methods, such as the gasometric
method (that is: by supplying pure oxygen from a
reservoir or by using electrolysis) or the paramagnetic
method.
Following the gasometric method, in a closed
headspace the change of pressure or change of volume is
related to oxygen consumption in the liquid phase. This
149
information can be used to activate an oxygen production
system, based on an oxygen bottle or an electrolytic cell,
and the oxygen flow or electrical current can be
converted to the respiration rate. In fact the resulting
oxygen supply serves as the aeration, i.e. flowing gas.
Change of pressure can be measured by a pressure sensor.
No documentation exists that describes the use of volume
change to activate oxygen supply. Obviously the
gasometric method is based on measurement of the
oxygen transfer [kLa · (S*O2 -SO2 )] and not on the
measurement of the oxygen concentration in the gas
phase. The gas phase oxygen concentration must be
assumed constant (i.e. dCO2 / dt = 0). In any case the
gasometric method requires a CO2 absorption
arrangement in order to remove CO2 from the gas. CO2
in the gas that is produced during biodegradation and is
released to the gas phase from the liquid phase, would
otherwise interfere with the measurement of pressure or
volume change. Typically one mole of CO2 is produced
per mole of O2 in aerobic respiration, which by definition
means that no gas pressure change will occur.
Alternatively, the oxygen concentration in the gas
phase may be measured directly, which eliminates the
need to capture CO2 and allows continuous aeration of
the biomass with air. However, in addition to the gaseous
oxygen concentration, DO concentration in the liquid
phase must be measured because besides the mass
balance in the gas phase, the mass balance in the liquid
phase also needs to be considered. Likewise, the flow rate
of the gas needs to be measured, for example with a mass
flow controller.
Oxygen is one of the few gases that show
paramagnetic characteristics, so it can be measured
quantitatively in a gaseous mixture by using the
paramagnetic method. The method is based on the
change in a magnetic field as a result of the presence of
oxygen, and this change is proportional to the
concentration of gaseous oxygen.
Another method to measure oxygen concentration in
the gas phase is by using a mass spectrometer. Using this
more expensive equipment has the advantage that other
gases may also be measured, which then is usually
termed off-gas analysis. Especially CO2 may be
measured in the gas phase because it is a useful indicator
of biomass activity under all the redox conditions.
However, if CO2 in the off-gas is measured, and because
CO2 is associated with the carbonate system, extra
equipment will be needed to measure the pH and
bicarbonate concentration in the liquid phase (Pratt et al.,
2003). Note that a mass spectrometer is expensive
150
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
equipment that also requires special (calibration) gases,
which makes this method more an advanced laboratory
tool rather than a field application. An alternative may be
a less expensive infrared CO2 analyser.
Mass
spectrometer
Vent
Capillary
4-way
valve
Ice bath
pH
Vacuum pump
Dosing pumps
Dryer
MFC
Bioreactor
Acid/Base
He
DO
Water
bath
Vent
MFC
MFC
Feed gas
(eg. O2, CO2, Ar)
MFM
MFC
Figure 3.24 Scheme of a GFS-based respirometer setup (Pratt et al., 2003).
Figure 3.24 shows an example of a practical
implementation of the GFS principle. This respirometer
is based on off-gas analysis using mass spectrometry and
is integrated with a titration unit to account for the
interaction of CO2 production and evolution with the
acid/base buffering systems in the liquid phase.
The respirometer was used to examine the two-step
nitrification process, more specifically nitrite
accumulation in wastewater treatment systems operated
under varying environmental conditions, i.e. pH and DO
concentrations (Gapes et al., 2003).
3.4 WASTEWATER CHARACTERIZATION
Several methods have been developed and applied for the
characterization of wastewater, both in terms of pollutant
load characterization and in terms of toxicity assessment.
In what follows first the different respirometric methods
for evaluation of biochemical oxygen demand will be
described, followed by respirometric toxicity tests and,
finally, an overview of the methods for wastewater
fractionation.
3.4.1 Biomethane potential (BMP)
3.4.1.1 Purpose
A biomethane potential (BMP) test is performed when
the methane yield of a substrate needs to be known, e.g.
when a business case for an anaerobic digester is being
developed. Also, the methane production over time can
be of interest to optimize the solids retention time of the
digester or for the dimensioning of the biogas handling
equipment.
3.4.1.2 General
The BMP test is performed to test what the methane
potential of a sample is. BMP determines, to a certain
extent, both the design and economic details of a biogas
plant (Angelidaki et al., 2009). But the BMP test can also
be done to evaluate the performance of the biomass. For
example, the methane production rate can be used to
estimate the hydrolytic activity of the inoculum. BMP
tests are often used in literature but there is a large variety
in protocols. There have been some efforts to propose a
standard (Angelidaki et al., 2009).
RESPIROMETRY
151
Figure 3.25 depicts a typical result of a BMP test. In
this case the methane production over time of coarsely
filtered sewage is displayed.
450
accumulated during the BMP tests. This means that the
acidification and VFA consumption by methanogenic
consortia need to be in equilibrium. The ratio between
inoculum and substrate is therefore important. For
biomass from a digester the following ratio can be
applied, which will in most cases not lead to a net VFA
production:
Methane production (N-mL gVS-1)
400
2≥
350
VSSinoculum
VSsubstrate
Eq. 3.15
300
Notice that these are masses applied in the test,
expressed in grams of VS or VSS, not concentrations. To
obtain the BMP of the substrate, the background
production of methane from the inoculum (a blank
without substrate) needs to be measured in parallel and
subtracted from the produced gas of the inoculumsubstrate mixtures.
250
200
150
100
50
0
0
5
10
15
20
25
30
Time (d)
Figure 3.25 Results of the BMP test: methane production over time from
coarsely-filtered sewage. Measurements were carried out in triplicate. The
red line represents solids from the filter belt; green, solids from the drum
filter, and blue, a mix of both solids (Kooijman, 2015; unpublished data)
3.4.1.3 Test execution
In BMP tests, the amount of inoculum (required for
anaerobic digestion) is usually measured as volatile
suspended solids (VSS). Volatile solids (VS) is a possible
alternative. The difference between these methods is that
VS considers all the volatile fractions in the inoculum,
whereas VSS only considers solids larger than a certain
mesh, separated by filtration. VSS is preferred over VS
because viable biomass is not expected to pass through
the filter in the VSS measurement and thus VSS is more
closely correlated to viable biomass than VS; see also
Section 3.5.1. However VS measurement can also be
chosen because of its accuracy and simplicity of
measurement and generally low dissolved volatile solids
concentration in inoculum, compared with the volatile
particulates. For VS or VSS measurement, most
researchers use the standard method (APHA et al., 2012).
In BMP tests, the way to measure the substrate
depends on the purpose of the test and on the form of the
substrate. For wastewater biomass usually VS is used.
Liquid substrate such as wastewater can be quantified by
COD. When the BMP in conventional anaerobic
digestion is investigated, it is important that no VFAs are
To perform the BMP test, the reaction bottles need to
be incubated at the desired temperature. The standard of
mesophilic temperature is 35 °C. The temperature can be
kept constant by using a water bath. Stirring has to be
performed by impellers in the reaction bottle. When
using biomass with high viscosity in BMP tests,
insufficient stirring may negatively affect the digestion
rates and sometimes impellers in small test bottles (200400 mL) do not meet the stirring requirement. A solution
for this can be to incubate the reaction bottles in an
incubator shaker. The duration of the BMP tests depends
on their purpose and the characteristics of the substrate
but it is most common to use a run time of 30 days.
3.4.1.4 Data processing
In many cases anaerobic respirometric tests aim at
measuring the BMP of a substrate. As shown in Figure
3.25, the BMP depends on the time of digestion, similar
to the analysis of BOD (Section 3.4.2). The BMP is also
usually expressed per gram of VS. For waste-activated
sludge this is between 150-200 N-mL gVS-1. For primary
sludge this is 300-400 N-mL g VS-1. Since the hydrolysis
is often the rate-limiting step in anaerobic digestion
(Eastman and Ferguson, 1981), the methane production
rate is directly related to the hydrolysis rate and thus the
slope of a respirogram as shown in Figure 3.25 is a direct
measure of the hydrolysis rate at that point in time.
3.4.1.5 Recommendations
• Pressure and temperature correction
In order to determine the amount of methane that is
produced in an experiment, the volume in combination
152
with the pressure and temperature is required at each time
instant during a test. Typically the amount of produced
gas is expressed under standard conditions (usually
273.15 K, 0 °C and 1,013.25 mbar, 1 atm).
• Methane diffusion
Methane molecules are known to be able to diffuse
through plastics such as silicon. Therefore it is crucial in
the design of a BMP (or Specific Methanogenic Activity,
SMA, Section 3.5.2) test that the materials that are in
contact with the biogas, have poor diffusivity for
methane.
• pH indicator dye for the scrubbing solution
It is crucial that the liquid in the scrubber bottles and in
the measurement device as depicted in Figure 3.13 and
Figure 3.15, respectively, has a high (> 9) pH such that
CO2 and H2S are absorbed by the liquid. A lower pH will
result in erroneous measurements. To ensure that the
scrubbing liquid is not saturated, methylene blue dye can
be added, which turns the liquid blue when pH > 9
making it possible to visually verify the effectiveness of
the scrubber solution.
• Inoculum activity
The activity of methanogens is very prone to temperature
differences. Especially when performing SMA tests, it is
important that the activity of the methanogens is high
from the beginning of the experiments. Therefore it is
strongly advised to store the inoculum at 35 °C for 24 h
prior to the experiment. In this way, in a BMP (or SMA)
test performed at 35 °C, there will be no temperature
shock for the methanogens and the BMP (or SMA) will
not be affected by temperature.
• Micro and macro nutrients
During a BMP (or SMA) test there may be a lack of micro
and macro nutrients, which may affect the conversion
performance. If insufficient nutrients are present they
should be added. In literature there are various
suggestions for nutrient solutions for anaerobic digestion
tests (Angelidaki et al., 2009; Zhang et al., 2014).
• Oxygen inhibition
When water is in equilibrium with air, the concentration
of oxygen will be around 9 mg L-1 at room temperature
and sea level. Oxygen is known to inhibit methanogens.
Also oxygen will ‘consume’ COD in a mixture.
Therefore, it can be desirable (especially in SMA tests
where the rates of methanogens are measured) to remove
the oxygen from the substrate solution before mixing
with biomass. Flushing with N2 gas is commonly applied.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Very gently bubbling N2 gas for ~ 60 sec through the
substrate solution will remove oxygen.
• Gas tightness
Prior to the BMP test, the system of gas production and
gas measurement should be verified to be gas tight. This
can be done by injecting a known quantity of air into the
tube connected to the scrubber and gas flow meter to
ensure that everything is mounted correctly and gas tight.
The amount of air measured and the amount of air
injected must be equal. If not, there is probably a leak.
• Alkaline scrubbers
Alkaline scrubbers are typically used to remove acid
components from gas. Other compounds such as NH3 and
H2 are not removed in alkaline scrubbers. NH3 is often
present in significant amounts when the pH of the
digestate is high (close to 9). When there is a severe
acidification in the digester, H2 may be formed and thus
be present in larger quantities in the biogas. To avoid
substantial variations in the pH of the mixture during the
assay, a phosphorus buffer solution may be added to the
mixture.
3.4.2 Biochemical oxygen demand (BOD)
3.4.2.1 Purpose
The biochemical oxygen demand (BOD) test is
performed to assess the biodegradable organic matter
concentration of a water sample, e.g. to design a
wastewater treatment plant (WWTP) or to evaluate its
performance in terms of organic matter removal. Thanks
to its sensitivity, it is also used to evaluate the organic
matter concentration of the receiving waters.
3.4.2.2 General
The determination of the BOD of wastewaters, effluents,
and polluted receiving waters is based on a test that
measures the bacterial consumption of oxygen during a
specified incubation time. The test quantifies the
biochemical
degradation
of organic
material
(carbonaceous biochemical oxygen demand - CBOD) but
it will also include the oxygen used to oxidize inorganic
material such as sulphides and ferrous iron. Unless a
nitrification inhibitor is added in the test, it may also
measure the amount of oxygen used to oxidize reduced
forms of nitrogen (nitrogenous biochemical oxygen
demand - NBOD). To make clear what is meant, the term
‘total BOD’ (BODt) is used when no nitrification
inhibitor is added, i.e. the sum of CBOD and NBOD.
RESPIROMETRY
153
Normally the incubation time is limited to five days,
leading to the traditional BOD5. However, tests can be
conducted over other incubation times, e.g. seven days to
facilitate lab organization, or 28, 60 up to 90 days of
incubation to determine the so-called ultimate BOD
(also: UBOD, BOD ∞ or BODU). This measures the
oxygen required for the total degradation of organic
material (the ultimate carbonaceous demand) and/or the
oxygen to oxidize reduced nitrogen compounds (the
ultimate nitrogenous demand).
Measurements that include NBOD are not generally
useful for assessing the oxygen demand associated with
organic material. In fact, NBOD can be estimated directly
from nitrifiable nitrogen (ammonia or total Kjeldahl
nitrogen) and CBOD can be estimated by subtracting the
theoretical equivalent of the reduced nitrogen oxidation
from the uninhibited test results. However, this method is
cumbersome and is subject to considerable error. The
chemical inhibition of nitrification provides a more direct
and more reliable measure of CBOD.
700
BODt (g m-3)
600
BOD5= 390 g m-3
400
300
200
100
0
0
5
10
Time (d)
15
In what follows the first test approach will be
presented and subsequently the methods using oxygen
supply from a gas phase are discussed.
• BOD test with a LSS respirometer
The method consists of completely filling an airtight
bottle with a water sample and incubating it in the dark
(to prevent photosynthesis) at 20 ± 0.1 °C for a specified
number of days (5, 7, 28, 60, 90 days). Dissolved oxygen
concentration is measured initially and after incubation,
and the BOD is computed from the difference between
the initial and final DO. Because the initial DO
concentration is determined shortly after the dilution is
made, all the oxygen uptake occurring after this
measurement is included in the BOD measurement. The
bottles are typically 300 mL with a ground-glass stopper
and a flared mouth. Bottles should be cleaned carefully
with detergent and rinsed thoroughly (to eliminate any
detergent from the bottle).
Dilution
BOD∞= 600 g m-3
500
In other methods oxygen is supplied continuously
from a gas phase present in the bottle (LFS respirometer)
and the oxygen consumed is monitored.
20
Figure 3.26 The BODt curve for mixed municipal and slaughterhouse
wastewaters (Henze et al., 1995).
Since the only source of oxygen is that initially present,
only limited oxidation can take place because DO
concentration should never decrease below 2 mg O2 L-1
to prevent oxygen limitation. This therefore limits sample
BOD concentrations to about maximum 7 mg L-1.
Whereas this may be adequate for effluent and receiving
water samples, wastewaters will have to be considerably
diluted in this closed bottle test. Dilution comes however
with problems since bacterial growth requires nutrients
such as nitrogen, phosphorus, and trace metals (Mg, Ca,
Fe). Without these the biodegradation of the pollutants
may be limited, leading to underestimation of the BOD.
Also, buffering may be needed to ensure that the pH of
the incubated sample remains in a range suitable for
bacterial growth. Obviously the dilution water used
should not contain biodegradable matter.
Seeding
3.4.2.3 Test execution
There are basically two measuring principles for BOD.
One employs a closed bottle (LSS respirometer) and the
only oxygen available for oxidation of the organic matter
is the oxygen dissolved in the (diluted) sample at the
beginning of the test.
Since the test depends on bacterial activity to degrade the
organic matter present in the sample, it is essential that a
population of microorganisms is present that is capable
of oxidizing the biodegradable matter in the sample.
Domestic wastewater, non-disinfected effluents from
biological WWTP and surface waters subjected to
wastewater discharge contain satisfactory microbial
populations. Some waters, e.g. industrial wastewaters,
may require a ‘seed’ to initiate biodegradation. Such a
seed can be obtained from the biomass or effluent of a
154
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
WWTP, but since nitrifiers may be present in such seed,
it is recommended to apply a nitrification inhibitor in the
test to ensure proper CBOD test results. In some cases the
pollutants may require organisms other than the ones
present in domestic WWTP and then it is recommended
to seed with bacteria obtained from a plant that is subject
to this waste or from receiving water downstream the
discharge point.
DO measurement
Dissolved oxygen can be measured with the azide
modification of the titrimetric iodometric method or by
using a well-calibrated DO electrode. For the
measurement of the ultimate BOD over extended
incubation periods, only the DO electrode measurement
approach is recommended because DO must be measured
intermittently during the incubation (intervals of 2 to
maximum 5 days, minimum 6 to 8 values).
Blank
Both the dilution water and seed may affect the result of
the BOD test, e.g. by introducing organic matter into the
bottle. In fact four situations may occur as illustrated in
the scheme below:
Data processing
The calculation of the BOD of the sample is as follows
(Figure 3.27):
BOD =
Seed
x
x
1
2
3
4
Dilution
x
x
Since the test quality could be affected by dilution, it
should be ensured by performing a blank BOD test in
which the same amount of seed as in the sample test is
added to a bottle filled with water for dilution (Figure
3.27).
D1 ‒ D2 ) ‒ (B1 ‒ B2 )
P
Eq. 3.16
Where, D1 is the DO concentration of the diluted
sample immediately after preparation (mg L-1), D2 is the
DO concentration of the diluted sample at the end of the
incubation period (mg L-1), B1 is the DO concentration of
the blank immediately after preparation (mg L-1), B2 is
the DO concentration of the blank at the end of the
incubation period (mg L-1), and P is the decimal
volumetric fraction of the sample used.
Note that in fact the mass balance of the closed bottle
test is written as (the LSS respirometric principle):
dSO2
= ‒ rO2
dt
B1
D1
DO concentration
Blank
B2
From which follows that, after integration,
BOD = SO2,t0 ‒ SO2,tfin =
Sample
D2
Incubation
time
Time (d)
Blank
Sample
Dilution sample
+ seed
Figure 3.27 A typical BOD closed bottle test result with a sample and a
blank.
tfin
rO2
t0
t) · dt
Eq. 3.18
Showing that the BOD is nothing else but the area
under a so-called respirogram, i.e. a time series of
respiration rate data.
For the determination of the ultimate BOD, the
following first order equation should be adjusted to the
time series of DO depletion data (Figure 3.28):
BODt = BODU (1 ‒ e-kt )
Dilution water
+ seed
Eq. 3.17
Eq. 3.19
Where, BODt is the oxygen uptake measured at time
t (mg L-1), BODU is the ultimate BOD (mg L-1), and k is
the first order oxygen uptake rate coefficient (d-1).
RESPIROMETRY
155
250
BODU
vigorously shaking a partially filled BOD bottle or by
aerating it with clean compressed air.
-1
BODt (mg O2 L )
200
150
100
50
0
0
5
10
Time (d)
15
20
Figure 3.28 Measured BOD of the sample over a period of 14 days and fit
of a first-order equation (Weijers, 2000).
Equation adjustment should preferably be performed
using nonlinear regression, e.g. by using the Solver
function in Excel to minimize the sum of squared errors
between the measured time series of oxygen uptake and
the model predictions of BODt for certain BODU and k
values. Note that this approach not only yields BODU but
also k. The latter provides information on the degradation
rate of the organic matter. Note that a first-order model
may not always be the best choice. Much better fits can
often be obtained with alternative kinetic models, in
particular consisting of a sum of two or more first-order
models.
Recommendations
When the DO concentration at the end of the test is below
1 mg L-1 or the DO depletion is less than 2 mg L-1, the
test should be carried out again, but with a higher or
lower dilution, respectively.
To verify whether the BOD test has been performed
well, a test check can be made with a solution with known
BOD. The recommended solution is a standard mixture
of 150 mg L-1 glutamic acid and 150 mg L-1 glucose. A 2
% dilution of this stock solution should lead to a BOD5
of approximately 200 ± 30 mg L-1. A BODU of 308 mg
L-1 can be anticipated (APHA, 2012).
Sample pre-treatment may be necessary. The
temperature and pH of the sample may have to be
adjusted to 20 °C and 6.5 < pH < 7.5 before dilution. If
the sample has been chlorinated, it should be
dechlorinated (by adding Na2SO3) and a seed should
certainly be used. Samples supersaturated with oxygen
(above 9 mg L-1 at 20 °C) may be encountered when the
sample was cold or there has been photosynthetic
activity. In such cases deoxygenation is necessary by
Nitrification can be inhibited using multiple
chemicals, e.g. nitrapyrin, allylthiourea (ATU), or 2chloro-6-(trichloromethyl)pyridine (TCMP). While
recommended concentrations normally lead to adequate
inhibition, adaptation to these chemicals has been
reported and they can also be degraded during the test,
allowing nitrification to start up at a later stage in the test.
It is therefore recommended to check at the end of the test
whether nitrite and nitrate have been formed.
(Field) test kits are available to measure BOD without
lab equipment. Standard nutrients and seed bacteria are
supplied, and DO is measured based on the photometric
method.
• A BOD test with a GFS respirometer
To solve the problem of a limited amount of dissolved
oxygen available in the closed bottle test, BOD test
equipment has been developed. The oxygen is supplied
to the liquid within this test equipment, thus enabling the
degradation of organic matter. In this way, dilution may
not be necessary or reduced significantly as the BOD
measurement range can be extended significantly. The
basic equipment consists of a bottle in which a gas phase
is provided that can supply oxygen as it is consumed in
the tested sample. Either the volume of gas contains
sufficient oxygen to allow completion of the oxidation of
the biodegradable matter in the sample, or the equipment
is able to replenish the gas phase with fresh oxygen from
an external source. In this way, besides BOD, the oxygen
uptake can also be measured more or less continuously
over time, which inherently provides the possibility to
calculate the respiration rate over time.
There is various equipment that works according to
the GFS principle. Basic principles include manometric
respirometers that monitor the pressure change in the gas
phase above the liquid as oxygen is consumed.
Interference by CO2 that is produced during
biodegradation and released from the liquid is dealt with
by capturing the CO2 in an alkaline (KOH) solution or
granules integrated in the equipment. Using the ideal gas
law the recorded pressure drop can be translated into
oxygen uptake. Equipment using a pressure transducer,
simple calculation logic and a data logger for the BOD
time series is commercially available. Volumetric
respirometers record the gas volume reduction (at
constant pressure) as oxygen gets depleted in the gas
phase. Again, CO2 scrubbing with alkaline
solution/granules is required. Electrolytic respirometers
156
use the principle of constant volume and gas pressure to
activate the electrolytic production of oxygen and
maintain the oxygen concentration in the gas phase
constant (Figure 3.29). CO2 emitted from the liquid
would again interfere with the constant pressure principle
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
and must be eliminated from the gas phase. Alternatively,
oxygen may be supplied, and its flow be measured, from
a pure oxygen source (e.g. gas bottle) to maintain
pressure in the airtight bottle.
Figure 3.29 An electrolytic respirometer for BOD analysis (SELUTEC GmbH).
If no external oxygen supply is available, as in the
respirometer described earlier (“BOD test with a LSS
respirometer”), one should pay attention to the overall
oxygen consumption that will be exerted by the sample.
This should not exceed the amount of oxygen present in
the gas phase above the liquid as this would lead to
oxygen limitation and thus erroneous results. The volume
of sample to be added in the bottle will thus depend on
its BOD content and respirometer manuals will typically
provide a table with volumes to be added for different
BOD ranges. In any case, methods without oxygen
supply suffer from decreasing DO concentration in the
liquid, which may influence the oxidation rate during the
test.
The instruments allow readings of the BOD as
frequent as every 15 min to every 6 h. High frequency
data collection may help in interpreting the results in
terms of degradation kinetics or allow a better model to
fit to reduce the influence of measurement noise or get a
more reliable estimate of the ultimate BOD.
Recommendations
Interference by gases other than CO2 may lead to
erroneous results, but is not often reported. Temperature
variations may also affect pressure and volume
measurement as they affect overall pressure. In addition,
atmospheric pressure changes may affect the
measurement in some respirometers.
An oxygen demand as small as 0.1 mg L-1 can be
detected, but the test precision will depend on the total
amount of oxygen consumed, the precision of the
pressure or volume measurement and the effect of
temperature and atmospheric pressure variations.
Since oxygen must be transferred from the gas phase
to the liquid phase, one must be careful that the oxygen
RESPIROMETRY
uptake rate does not exceed the mass transfer rate by too
much as this would lead to oxygen limitation of the
biodegradation process and thus errors in the BOD result.
The oxygen transfer is mostly dependent on the mixing
conditions in these respirometers, thus limiting the
oxygen uptake rates to 10 mg L-1 h-1 for low mixing
devices and to 100 mg L-1 h-1 if high intensity mixing is
provided. If the uptake rate exceeds the supply rate of the
respirometer at hand, one can dilute the sample such that
the uptake rate decreases to acceptable values. The
dilution water composition must be checked as
mentioned above.
Note that the BOD test is a respirometric test because
it measures consumption of the terminal electron
acceptor O2, although it generally does not measure the
respiration rate. In fact BOD is the cumulative oxygen
consumption that may be obtained by integrating the
respiration rate over a certain incubation period. As noted
above, the LFS respirometry inherently provides the
possibility to calculate the respiration rate over time.
Also note that BOD can be measured in similar ways
by using other electron acceptors, such as nitrate.
Seeding and nutrient additions may be required for
particular wastewater samples if competent biomass is
not present or the wastewater is not balanced in terms of
nutrients versus organic material and may thus be subject
to limitations of bacterial growth and thus biased BOD
results.
3.4.3 Short-term biochemical oxygen
demand (BODst)
Temporal variation of the wastewater composition can be
readily characterized by using chemical methods such as
COD and TOC analysis. These methods can provide data
at high measuring frequency (e.g. on an hourly basis) but
they do not provide information on the bio-treatability of
the pollutants. Traditional methods that rely on the
monitoring of the biodegradation of the pollutants to
obtain an indication of the treatability such as the
aforementioned BOD5 method are clearly unsuitable to
provide such high-frequency information due to the large
time delay between the sample introduction and
measurement result. However, the principle of
monitoring the oxygen uptake for assessment of the
treatability and organic matter concentration of a
wastewater is a very powerful one, because most
wastewater treatment processes rely on aerobic
degradation of the organic matter. Therefore, methods
157
have been proposed to decrease the response time of
these biologically mediated methods to such a level that
their application in high-frequency monitoring becomes
possible.
The short-term biochemical oxygen demand (BODst)
is defined as the amount of oxygen consumed for
biodegradation of readily biodegradable organic matter
per volume of wastewater.
In the traditional BOD5 test (Section 3.4.2), a small
amount of biomass is added to a large wastewater sample
(typically the initial substrate to biomass ratio S0/X0 is set
to be between 10/1 and 100/1 mg BOD5 mg MLVSS-1).
As a result, substantial growth has to occur before the
available pollutants are degraded and possibly a lag phase
may occur where adaptation of the sludge to the
pollutants takes place. To speed up the response time (to
within an hour), the techniques for BODst determination
are based on a low S0/X0 ratio (typically 1/20 to 1/200 mg
BOD5 mg MLVSS-1). These conditions are obtained by
the addition of a small aliquot of wastewater to the
activated sludge present in the test vessel. In this way the
degradation time can be reduced considerably (to often
less than one hour) and no significant growth of biomass
is to be expected given the relatively low amount of
organic matter added.
Because of this short test time, it is evident that matter
that biodegrades slowly in the wastewater sample will not
be degraded, i.e. only the readily biodegradable fraction
is measured. There is also no time for adaptation of the
biomass to any new organic component that may be
present in the wastewater and to which the biomass is
exposed for the first time. In addition there is an
important issue of not using biomass from an enhanced
biological sludge removal (EBPR) plant. This biomass
contains phosphorus accumulating organisms (PAO) that
can store volatile fatty acids (VFA). When the VFA
fraction of the readily biodegradable matter is taken up
for storage it will not be oxidized and hence it will not be
measured in a respirometric test. The reader is also
referred to sections 3.5.3.2 and 3.5.3.3 (Figure 3.48).
However, being able to measure the readily
biodegradable matter is very relevant information in
particular for anticipating and optimizing the
performance of denitrification and enhanced biological
phosphorus removal. Note that, because a respirometric
test can be used to distinguish between slowly and readily
biodegradable matter, this provides a basis for the
assessment of wastewater fractions in the context of a
mathematical model (Section 3.4.5).
158
3.4.3.1 Test execution
The BODst test is performed by first bringing a volume
of activated sludge in the thermostated (e.g. 20 °C) and
aerated test vessel and letting it come to a stable state
characterized by what is called endogenous respiration.
Endogenous respiration is defined as the state the
biomass gets in when there is no more external substrate
available in the activated sludge. The biomass is respiring
its own reserve materials or is respiring the lysis products
of dead biomass. Reaching this state may take between a
few hours and a day, depending on how loaded the
activated sludge is with slowly biodegradable
(hydrolysable) matter. Typical endogenous respiration
rates are between 2 and 10 mg O2 L-1 h-1. Once the
endogenous state has been reached a pulse of wastewater
is injected, the amount of which is calculated to lead to
the desired S0/X0-ratio of approximately 1/20 to 1/200
mg BOD5 mg MLVSS-1. As soon as the substrate
becomes available to the biomass, biodegradation starts
and the oxygen uptake rate increases quickly to a
maximum rate that is determined by the activity of the
biomass and the degradation rate of the substrate. If
nitrifiers are present in the sludge and the wastewater
contains nitrifiable nitrogen (ammonia and organic
nitrogen that is ammonified rapidly) this exogenous
respiration rate also includes the oxygen uptake rate for
nitrification. The substrates will gradually get exhausted
and the oxygen uptake rate will gradually decrease to
eventually reach the endogenous rate. The substrate
pulse-induced curve of oxygen uptake rates is called a
respirogram, a schematic of which is given in Figure
3.30.
rO2 (mg O2 L-1 h-1 )
100
Sample
80
60
40
rO2,endo
20
0
0.00
0.25
0.5
Time (h)
0.75
1.00
Figure 3.30 Schematic of a respirogram. After biomass has undergone
endogenous respiration (no external substrate is present), a sample
containing organic matter is injected. The exogenous respiration starts and
continues until all the substrate has been removed, and respiration returns
to the endogenous rate level.
Figure 3.31 illustrates that exogenous heterotrophic
and autotrophic respiration are independent and their
respirograms can be superimposed. Three respirograms
are shown, one with only COD addition as acetate
showing heterotrophic activity, one with ammonia
addition showing nitrifier activity and one with the
COD/N mixture.
1.2
C
N
C+N
1.0
rO2,exo (mg O2 L-1 min-1)
Due to the increased respiration resulting from the
high degradative capacity available in the test vessel,
problems may arise to meet the oxygen requirements of
the sludge. Respirometric methods that are based on
measurement of the decrease of the initial amount of
oxygen present in the biomass are severely restricted
because of the risk of oxygen limitation. As a result the
dynamic concentration range of such methods is small,
with maximum sample additions of 5 mg BODst L-1
activated sludge. Obviously, aeration of the sludge,
available in many respirometric equipment, solves this
problem. However, the intensity of aeration must be
sufficient. Still, since in many cases the dissolved oxygen
is measured in the test vessel, it can easily be checked
whether the oxygen supply has been sufficient (that is:
DO should always be above 2 mg L-1) during the BODst
test. Initial BODst concentrations in the test vessel can
easily reach 100 mg BODst L-1 in most respirometers.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
0.8
0.6
0.4
0.2
0.0
0
5
10
15
20
25
30
Time (min)
Figure 3.31 Illustration of adding together the exogenous heterotrophic
and autotrophic respiration rates. The square symbols represent a
respirogram with the addition of 20 mg COD L-1 of acetate, the circles
represent the respiration rate due to nitrification after the addition of 2.5
mg N L-1 and the line is the respirogram of the mixture of 20 mg COD L-1 and
2.5 mg N L-1 (Kong et al., 1996)
Figure 3.32, Figure 3.33 and Figure 3.34 show some
typical respirograms, some of them only showing the
RESPIROMETRY
159
exogenous respiration rates that, following some
respirometric principles, can be directly calculated from
the DO data.
0.5
rO2 (mg O2 L-1 min-1)
1.0
r O2,exo (mg O2 L-1 min-1)
The presence of respiration due to nitrification is
noteworthy in the respirogram in Figure 3.32. Indeed,
when a wastewater sample supplemented with extra
ammonium is injected, one can clearly observe additional
exogenous respiration due to respiration for nitrification
in the absence of nitrification inhibition.
respiration rate is observed. This sudden decrease is due
to the complete removal of ammonium from the activated
sludge, after which only hydrolysable organic matter is
being oxidized, hence the typical exponential decrease in
respiration rate due to the first-order nature of hydrolysis.
0.4
Wastewater + extra ammonium
0.3
0.8
0.6
0.4
0.2
0.0
0
0.2
20
40
60
80
100
120
Time (min)
Wastewater
0.1
Figure 3.33 An example of respirograms obtained from BODst tests with
municipal wastewater (Spanjers and Vanrolleghem, 1995).
0.0
0
20
40
60
80
100
120
Time (min)
Figure 3.32 Example of respirograms obtained from BOD tests with
wastewater and wastewater mixed with an additional amount of
ammonia (Petersen et al., 2002a).
The final example presents a respirogram of an
industrial wastewater (Figure 3.34).
When two experiments are performed, one in the
presence of a nitrification inhibitor and one without an
inhibitor, one can estimate the concentration of
ammonium in the wastewater by making the difference
between the BODst of both samples. This oxygen demand
divided by 4.33 mg O2 mg N-1 enables the ammonium
concentration in the wastewater sample to be obtained
(provided that full nitrification to nitrate has been
achieved). Moreover, since ammonification is normally
a very quick process, it is not only ammonium that can
be quantified like this, but in fact the nitrifiable nitrogen
in the wastewater.
Figure 3.33 shows an example of a respirogram for a
typical municipal wastewater sample added to biomass
without a nitrification inhibitor. In this graph only the
exogenous respiration rate is depicted. The interpretation
of the respirogram goes as follows (Spanjers and
Vanrolleghem, 1995). Starting from the right end of the
respirogram, the respiration rate increases gradually until
about 50 minutes at which point a sudden decrease in the
rO2,exo (mg O2 L-1 min-1)
1.0
0.8
0.6
0.4
0.2
0.0
0
10
20
30
40
50
Time (min)
Figure 3.34 The results of the OUR test with industrial wastewater
depicting three superimposed respirograms corresponding to the
degradation of three different solvents (Coen et al., 1998)
Again, only exogenous respiration rates are shown.
The respirogram shows in fact the superposition of three
respirograms. This could be explained by the presence of
three major solvents in the wastewater that were
160
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
degraded in parallel by the acclimated biomass. By
making three horizontal lines corresponding to the
plateaus, one can delineate the amounts of each of the
solvents by taking the corresponding area (see the
calculations below). Note that in Figure 3.34 horizontal
lines can be used because the measured rates correspond
with substrate saturation conditions, whereas in some
other cases (for example Figure 3.39) substrate limitation
occurs, i.e. the respiration rate decreases as the substrate
is exhausted, and sloped lines are used.
3.4.3.2 Calculations
Since BODst (mg O2 L-1) is defined as the amount of
oxygen consumed for biodegradation of pollutants, it can
be readily calculated by integration of the time series of
exogenous respiration rates rO2,exo (mg O2 L-1 h-1):
tfinal
BODst =
tpulse
rO2,exo t) · dt
Eq. 3.20
Where, tpulse is the time of pulse addition, and tfinal is
the time needed to return to the endogenous respiration
rate after the sample addition.
Evidently the BODst of the injected sample is
deduced from:
sample
BODst
=
VSludge + VSample
∙ BODst
VSample
Eq. 3.21
Where, VSample is the volume of sample and VSludge is
the volume of sludge in the test vessel prior to the sample
addition.
As explained before in the context of Figure 3.31, the
amount of nitrifiable nitrogen (Nnit in mg N L-1) can be
calculated from the part of the respirogram that can be
attributed to the respiration rate due to nitrification rNit
O2,exo
(mg O2 L-1 h-1):
NNit =
1
4.57 ‒ YANO
tfinal
tpulse
rNit
O2,exo t) · dt
Eq. 3.22
Where YANO is the nitrifier’s yield coefficient,
typically 0.24 mg COD mg N-1.
Similarly, the BODst attributed to different fractions
of organic matter present in a wastewater can be
calculated from the exogenous respiration rates that can
be attributed to the superimposed respiration rates riO2,exo
(mg O2 L-1 h-1), Figure 3.33:
BODist =
tfinal
tpulse
riO2,exo t) · dt
Eq. 3.23
3.4.4 Toxicity and inhibition
3.4.4.1 Purpose
Respirometric techniques have been frequently preferred
for the assessment of inhibitory and toxic effects of
substances or wastewater on biomass (Volskay and
Grady, 1990). Inhibition is the impairment of biological
function and is normally reversible. Toxicity is an
adverse effect on biological metabolism, normally an
irreversible inhibition (Batstone et al., 2002). In this
chapter we prefer to use the terms toxicity and toxicant
to cover both reversible and irreversible effects.
Although toxicity test results are often expressed in terms
of IC50 (the concentration which produces 50 %
inhibition of the respiration), the IC50 does not give
complete insight into the toxic effect of a chemical. In
order to predict the effects of the toxic compounds on the
removal of organic matter and nutrients or to design
mitigative actions in the biological treatment process, the
effects of a toxicant on the biodegradation kinetics should
be quantified.
3.4.4.2 Test execution
Both decreases in the endogenous and exogenous
respiration rates can be used to assess toxicity. If the
decrease in the endogenous respiration rate is used as an
indicator of toxicity, biomass is first brought to the
endogenous state, the toxicant is added and then the
reduction in endogenous respiration rate measured. If the
toxicant or toxic wastewater is biodegradable, exogenous
respiration may occur, interfering with the evaluation of
the reduction in the endogenous respiration rate. It has
also been found that biomass is less sensitive to toxicants
when it is in its endogenous state.
For toxicity assessment based on exogenous
respiration rates, respirograms (see Section 3.4.3.1) are
the basis. In an exogenous rate-based test, a reference
substrate (e.g. acetate for heterotrophic toxicity tests or
ammonia for nitrifier toxicity tests) is injected prior to
injection of a potentially toxic sample so as to assess the
reference respiration rate (i.e. biomass activity). After the
toxicant has been added and has had time to affect the
biomass, another pulse of reference substrate is injected
and the determined activity (e.g. the maximum
exogenous respiration rate) is compared to that obtained
RESPIROMETRY
161
prior to the toxicant injection. It is to be noted that the
time between the toxicant injection and reference
substrate injection may affect the level of toxicity
determined since longer-term exposure may lead to a
stronger effect or may also lead to biomass adaptation
(Figure 3.35). No clear indication is available on optimal
exposure times. Sometimes even stimulation of
respiration may occur, i.e. the maximum respiration rates
for reference substrate degradation after the addition of
the toxicant are higher than without the toxicant. This is
due to increased energy requirements by the biomass to
cope with the toxicant (e.g. energetic uncoupling in the
presence of benzoic acid as the toxicant). In any case it is
important to always report the applied time between the
injections of the toxicant and the reference substrate.
0.15
A)
0.10
3.4.4.3 Calculations
A toxicity severity level can then be deduced by
calculating the ratio in activity levels before and after the
toxicant addition.
Toxicity %) =
max
rmax
O2,exo before) ‒ rO2,exo after)
×100
rmax
O2,exo before)
Eq. 3.22
Obviously the toxicity severity level will depend on
the dose of toxicant applied. Dose-response curves are
obtained from a series of experiments in which the
toxicity response is measured at different doses, and they
allow the assessment of the IC50 (the half maximum
inhibitory concentration). Normally the testing protocol
to obtain the dose-effect relationship is to perform a
reference-toxicant-reference sequence at one dose,
replace the biomass with fresh biomass and perform
another reference-toxicant-reference sequence at a higher
dose. Typically the concentration of the dose is a multiple
of the previous dose. This set of reference-toxicantreference sequences is continued until the biomass is
completely inhibited.
0.05
Control
400 mgL-1 Chloroform
0.00
r O2 (mg O2 L-1 min-1)
0.15
B)
0.10
0.05
Control
42 mgL-1 4-Nitrophenol
0.00
0.15
C)
0.10
Control
100 mgL-1 4-Xylene
0.05
0.00
0
20
40
60
80
100 120
140 160
180
Contact time (min)
Figure 3.35 Typical data from respiration inhibition tests with different
contact times (A) Chloroform, (B) 4-Nitrophenol, and (C) 4-Xylene (Volskay
and Grady, 1990)
To speed up the experimentation the step of replacing
the biomass can be omitted (Kong et al., 1994), but then
the contact time is not under the control of the
experimenter. In Figure 3.36 an example of such fast
determination of the dose-effect relationship is given for
copper addition to biomass in increasing doses of 0, 2.5,
5.0 and 10.0 ppm. This leads to a copper exposure
concentration of 0, 2.5, 7.5 and finally 17.5 ppm, but each
with different contact times. The figure clearly
demonstrates that the maximum respiration rate for
degradation of acetate (the reference substrate in this
experiment) decreases rapidly. It is noteworthy that the
area under the curve remains the same, i.e. the BODst
(Section 3.4.3) remains the same since all the acetate is
still degraded, albeit at a lower rate. This also means that
the respirogram takes longer to complete.
From this experiment the IC50 can be visually
deduced to be approximately 7.5 ppm since the
maximum respiration rate at this copper concentration is
about half the maximum respiration rate in the absence
of copper (compare the first, at ~0.5 h, with the third
respirogram, at ~1.7 h).
162
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
-1
rO2,exo (mg O 2 L-1 min )
1.00
0.75
0.50
0.25
0.00
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Time (h)
Figure 3.36 Typical respirograms obtained from an ARIKA (Automated
Respiration Inhibition Kinetics Analysis) test with copper (Kong et al.,
1994). The first respirogram is obtained using pure acetate and the
successive respirograms are a series of results where a mixture of acetate
with geometrically increasing toxicant concentration (cumulative toxicant
concentration: 2.5, 7.5 and 17.5 ppm) was used.
1.4
Peak
2
1.2
A)
3
4
1.0
CN
5
-
6
0.8 1
7
0.6
rO2,exo (mg O2 L-1 min-1)
8
0.4
9
0.2
0.0
1.4
B)
1.2
Peak
2
1.0
Cu
2+
3
4
0.8
1
0.6
5
6
0.4
0.2
0.0
0
1
2
3
4
5
6
7
8
9
Time (h)
Figure 3.37 Exogenous respirograms with C and N (20 mg L-1and 2 mg L-1,
respectively) substrate addition with Cu2+ (A) and CN- (B) as toxicants. The
first peak is the one for pure acetate, the second one is the pure mixture of
C and N, followed by a series of mixtures of C and N substrate and toxicant
(cumulative Cu2+ concentrations of 2.5, 7.5, 17.5, 37.5, 77.5 and 157.5 ppm
and cumulative CN- concentrations of 0.013, 0.038, 0.088, 0.188, 0.388,
0.788 and 1.588 ppm). (Kong et al., 1996).
As previously mentioned, autotrophic and
heterotrophic respirograms can be superimposed and
toxicity testing towards heterotrophic and autotrophic
activity can thus be done in a single experiment with a
mixture of COD and N as the reference substrate. In
Figure 3.37 the impact of two classical toxicants, copper
and cyanide, on heterotrophic and autotrophic biomass
are assessed. These respirogram series with increasing
toxicant concentration show that nitrifiers are more
sensitive to cyanide than heterotrophs (IC50 is
approximately 0.2 ppm for nitrifiers, whereas it is about
1 ppm for heterotrophs). For copper the IC50 for acetate
oxidizers and IC50 for nitrifiers are approximately the
same as the whole respirogram decreases together.
3.4.4.4 Biodegradable toxicants
So far, only the toxicity of non-biodegradable substances
has been considered. However, using respirometry for
toxicity determination of biodegradable substances is
more involved. When the typical reference-toxicantreference sequence is used to determine toxicity, a
biodegradable toxicant will lead to a respirogram after
injection; thus this must be completed before the second
reference substrate pulse can be added. If the toxic impact
is reversible (i.e. inhibition), it may be that no toxicity
can be detected since the toxicant has been removed from
the activated sludge. However, contrary to this nonlasting, or temporary, toxicity, lasting toxicity may be
assessed with the reference-toxicant-reference sequence,
as explained above.
Respirometry can be very useful to study a number of
biodegradable toxicants that are inhibitory to their own
biodegradation through what is called substrate
inhibition, i.e. the substrate on which biomass is growing
is inhibitory to the biomass growth process. Nitrification
may also be self-inhibited when the ammonia
concentration is too high. Respirograms of such selfinhibitory substrates show an increasing respiration rate
as the concentration of the biodegradable toxicant
decreases with time and its inhibitory effect thus
diminishes.
Figure 3.38 illustrates how respirometry can explain
complex phenomena occurring in biological wastewater
exposed to toxicants (Gernaey et al., 1999).
RESPIROMETRY
163
-1
rO2 (mg O 2 L-1 min )
0.5
3.4.5 Wastewater fractionation
0.4
0.3
0.2
Phenol
degradation
finished
0.1
Nitrification
finished
0.0
0
20
40
60
80
100
120
140
160
180
Time (min)
Figure 3.38 The respirometric response of a mixed heterotrophicnitrifying biomass to the addition of a mixture of phenol (15 mg L-1) and
ammonium (5 mg N L-1). Only the exogenous respiration rate is shown
(Gernaey et al., 1999).
In this example, phenol and ammonia are added to a
mixed biomass consisting of nitrifying and heterotrophic
bacteria. Phenol is one of the best-known examples of a
toxic but biodegradable compound. In a preliminary
experiment it was confirmed that phenol could be
degraded by the biomass used in the experiments. What
is observed in the respirometric experiment depicted in
Figure 3.38 is that nitrification is inhibited by the phenol
that is gradually degraded by heterotrophs. However, the
phenol degradation itself is also inhibited by the toxic
effect of phenol on the heterotrophs, that is: selfinhibition. In the first phase, nitrification is inhibited but
the phenol is degraded, albeit slowly. Phenol degradation
speeds up after about 30 minutes and comes to an end
after about 50 minutes, as can be seen from the sharp
decrease in the rO2,exo profile. As soon as the phenol
degradation is completed, nitrification increases to the
same rate as before the phenol addition (a separate
experiment with ammonium addition may also be
performed to assess the effect on nitrification alone).
Because the inhibiting concentration depends on the
biomass’ origin and acclimatization, the toxicity of a
compound to acclimatized biomass may be interesting to
study. An example of acclimatization based on a
respirometric experiment, with a progressive increase in
the IC50 and inhibition coefficient, can be found in
Rezouga et al., (2009).
The use of dynamic models in activated sludge processes
has become more and more widespread and has become
a way of thinking and communicating about wastewater
treatment processes. The Activated Sludge Model No.1
(ASM1) presented by the IAWQ Task Group on
Mathematical Modelling for Design and Operation of
Biological Wastewater Treatment Processes (Henze et
al., 1987) is generally accepted as state-of-the-art, and is
used for simulation of wastewater treatment plants in
many studies. It will form the basis of this section, which
deals with the use of the respiration rate to obtain
wastewater fractions in the context of ASM1.
Before establishing the relationship between the
respiration rate and the fractions of the ASM1
components, it is essential to explain this model for the
heterotrophic process (Table 3.2). It is assumed that the
reader comprehends the Gujer matrix as depicted in
Table 3.2. In the mass balance of the heterotrophic
organisms XOHO (component 5, in short: c. 5), the
production of XOHO by aerobic growth (process, or
reaction 1, in short: r. 1) is counteracted by the loss of
XOHO by heterotrophic decay (r. 4). In this decay process
component XOHO (c. 5) is converted to component XCB
(c. 4). This production of XCB is counteracted by the loss
of XCB by hydrolysis (r. 7), leading to production of
component SB (c. 2). SS is used for heterotrophic growth
(r. 1) where it is converted to component XOHO (c. 5) at
the expense of component oxygen SO2 (c. 8), i.e.
respiration. A similar reasoning can be made for the
processes involving the soluble and particulate nitrogen
components (SNHx, SB,N and XCB,N) and autotrophic
(nitrifying) organisms (XANO). Anoxic growth (r. 2) with
nitrate SNOx (c. 9) as the terminal electron acceptor is also
considered.
Hence, the wastewater fractions that will be
quantified using aerobic or anoxic respirometry are the
biodegradable COD fractions: SB, XCB, the biomasses
potentially present in the wastewater: XOHO and XANO and
the nitrogen fractions: SNHx, SB,N and XCB,N. Other, inert
fractions can be determined as well using respirometric
data, but then additional chemical analyses of total and
soluble COD and total and soluble nitrogen fractions will
be needed. This will not be dealt with here, but can be
found in Vanrolleghem et al. (1999) and Petersen et al.
(2003).
Table 3.2 Activated Sludge Model No. 1: Gujer matrix (Henze et al., 1987).
164
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
RESPIROMETRY
rO2,tot =
1 − YOHO
4.57 − YANO
⋅ X OHO ⋅ μ OHO +
⋅ X ANO ⋅μ ANO
YOHO
YANO
Eq. 3.25
Where the specific growth rates μOHO and μANO are
functions of SB and SNHx, respectively (Henze et al.,
1987).
The concentrations of SB and SNHx, in turn, depend on
the rates at which XCB, SN,B and XCN,B are degraded
(Table 3.2). It is clear that all the independent processes
summarised in Table 3.2 eventually act on the mass
balance of oxygen (and nitrate if the same evaluation is
done for anoxic conditions).
There are two approaches for the assessment of
model wastewater fractions: direct methods focus on
specific fractions that can be directly evaluated from the
measured respiration rates (Ekama et al., 1986; Spanjers
et al., 1999), whereas optimisation methods use a (more
or less simplified) model that is fitted to the measured
data (Kappeler and Gujer, 1992; Wanner et al., 1992;
Spanjers and Vanrolleghem, 1995). In the latter,
numerical techniques are used to find values of the
unknown wastewater fractions that lead to the smallest
deviation between the model predicted and the measured
respiration rates. The optimisation methods for
estimation, using numerical techniques, will not be
discussed here, but reference is made to Chapter 5.
From the processes described in Table 3.2, it is clear
that the total respiration rate is affected by the
concentrations of all the biodegradable components and
that the cumulative oxygen consumption (i.e. the integral
of rO2) is a measure of the amount of components
degraded. Notice that because in direct methods integrals
of respiration rates are taken, the measuring frequency
and signal-to-noise ratios (measurement error compared
to measurement value) are not very critical for the
reliable assessment of the component concentrations.
This is in contrast to kinetic characterization
(Vanrolleghem et al., 1999 and Petersen et al., 2003),
where information is obtained from changes in
respiration rates (derivative of rO2). This implies a much
higher dependency on the parameter accuracy of the
respirometric measurement quality.
Inherent with respirometric tests for wastewater
characterization is the use of biomass: the assessment of
wastewater components is based on the respirometric
response of biomass to wastewater. Two important
aspects are associated with the use of biomass. The first
aspect is the amount of wastewater component with
respect to biomass (SO/XO ratio, Sections 3.4.1.3 and
3.4.3) that is used. Second, in the death-regeneration
concept adopted in ASM1, new SB, XCB, SNHx, SN,B and
XCN,B are continuously generated from decaying
biomass. Within this model it is therefore difficult to
distinguish between the components originating from the
wastewater and from the biomass itself. In fact the
transition between exogenous respiration and
endogenous respiration is gradual. The respirometric test
should thus be organised in such a way that these rates
can be distinguished. This is one of the most challenging
problems in respirometric characterization of wastewater
in the context of ASM1.
Figure 3.39 shows a respirogram collected in a batch
experiment where at the start wastewater is added to
endogenous sludge.
1.0
rO2,exo (mg O2 L-1 min-1)
The total respiration rate of biomass in contact with
wastewater is, according to the ASM1:
165
0.8
0.6
SB
0.4
S NHx
0.2
XC B
0.0
0
20
40
60
80
100
120
Time (min)
Figure 3.39 An exogenous respiration rate profile obtained after addition
of 0.7 L wastewater into 1.3 L activated sludge and fractionation according
to the procedure described by Spanjers and Vanrolleghem, 1995b.
In Figure 3.39 only the exogenous respiration rate is
presented, i.e. the endogenous respiration is subtracted
from the total respiration rate. A typical wastewater
respirogram shows an initial peak brought about by the
oxidation of readily biodegradable matter, followed by
one or more shoulders where successively other
components continue to be oxidised. The full area under
the respirogram represents the total of biodegradable
components (SB + XCB) / (1 ‒ YOHO) + (SNHx + SN,B +
XCN,B) / (4.57 ‒ YANO), as follows from the above
equation on total respiration rate. In Figure 3.39 three
166
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
3.4.5.1 Readily biodegradable substrate (SB)
The readily biodegradable substrate is presumably
composed of (simple/low molecular) soluble
compounds, such as volatile fatty acids, alcohols, etc.
(Henze et al., 1992). The characteristic of these
compounds is that they are degraded rapidly and hence
provoke a fast respirometric response.
A typical batch test for determination of SB (e.g.
Ekama et al., 1986) involves the addition of a wastewater
sample to endogenous sludge, and monitoring the
respiration rate until it returns to a level where it can be
reasonably assumed that the readily biodegradable
substrate is removed from the activated sludge. If other
processes occur that are also consuming oxygen (e.g.
endogenous and nitrification respiration rates), they must
be identified or assumed so that their respiration rates can
be subtracted from the total respiration rate (Figure 3.39).
Nitrification may be inhibited (Section 3.4.2). Only when
the other oxygen consuming processes are accounted for,
can the respiration rate related to oxidation of readily
SB
biodegradable substrate (rO2,exo
) be identified and used to
calculate the SB concentration in the wastewater.
The concentration of readily biodegradable substrate
initially present in the mixture of biomass and wastewater
can then be calculated as follows:
SB (0) =
t fin S
1
B
rO2,exo
( t ) dt
1 − YOHO 0
optimisation method may be a bit excessive, but for more
complex estimation tasks (next paragraph), the approach
becomes more straightforward than direct calculation
methods.
Notice that knowledge of the heterotrophic yield
coefficient YOHO is needed for the calculation of SB from
respiration rates. This stoichiometric coefficient is
always involved when oxygen consumption is converted
to substrate equivalents (see also next section and
Sections 3.4.5.2 – 3.4.5.6). The batch test described
above is also used to assess other ASM1 components and,
using the optimization method, kinetic parameters. This
explains its popularity in calibration procedures.
40
rO2 (mg O2 L-1 min-1)
substrate fractions can be discerned, corresponding to SB,
XCB and SNHx.
30
20
10
0
0
1
2
3
4
5
6
Time (h)
Figure 3.40 Respiration rates measured in a batch experiment (Kappeler
and Gujer, 1992) for the determination of readily biodegradable substrate
according to the method by Wentzel et al., 1995.
Eq. 3.26
The concentration of SB in the wastewater is then
easily calculated by taking into account the dilution. The
end point tfinal of the integration interval is the time instant
where SB is completely oxidised and where the
exogenous respiration rate for SB becomes zero. The
integral can easily be obtained by determining the area
under the curve, for instance by using a spreadsheet
program, also known as the graphical method. An
alternative to this direct method is the optimisation
method, as explained above. This consists of solving the
mass balance equations with a numerical integrator to
predict the exogenous respiration rates for SB in such a
batch experiment. Depending on the initial value SB(0)
given to the integration algorithm, the simulation will
result in a different predicted respirogram. One can
therefore search for the SB(0) value that gives the ‘best
fit’ to the measured data. For this simple application the
Another batch test (Wentzel et al., 1995) consists of
monitoring the respiration rate of unsettled sewage
without seed for a relatively long period (up to
approximately 20 h). A respirogram similar to the one
depicted in Figure 3.40 is obtained. SB is calculated from
the respiration rates observed between the start of the test
up to the sudden drop (due to depletion of SB), with
correction for the increasing endogenous respiration due
to the increase in biomass during the test. In addition to
YOHO, knowledge of the net growth rate is required,
which can be obtained from the same test.
Ekama et al. (1986) presented a method for
determining SB that involves respiration rate monitoring
in a completely mixed reactor operated under a daily
cyclic square-wave feed. It is hypothesised that the
sudden drop in respiration rate to a lower level, observed
upon termination of the feed (Figure 3.41), corresponds
RESPIROMETRY
167
uniquely to the SB that has entered via the influent.
Hence, the concentration of readily biodegradable
substrate in the wastewater can be calculated as:
SB =
Vreact ΔrO2,tot
⋅
Q ww 1 − YOHO
Eq. 3.27
Vanrolleghem, 1995). Alternatively, if the data of such
respirometric batch tests are evaluated using the
optimisation method to match the response of the model
to the data, the nitrification part can easily be extracted
from the respirogram (Spanjers and Vanrolleghem,
1995).
40
50
rO2 (mg O 2L-1 min-1)
rO2 (mg O 2 L-1 min-1)
40
∆ rO2
30
20
10
OFF
Feed ON
30
20
10
Feed OFF
0
0
0
2
4
6
8
10
12
14
16
18
20 22
24
0
Time (min)
Figure 3.41 Respiration rates (rO2) obtained using the experimental setup
of Ekama et al., (1986) and permitting the direct assessment of readily
biodegradable substrate SB.
3.4.5.2 Slowly biodegradable substrate (XCB)
It is presumed that slowly biodegradable substrate XCB,
sometimes also defined as particulate material, is
composed of (high-molecular) compounds ranging from
soluble to colloidal and particulate (Henze, 1992). The
common feature of these compounds is that they cannot
pass the cell membrane and must first undergo hydrolysis
to low-molecular compounds (SB) that can subsequently
be assimilated and oxidized. Because the rate of
hydrolysis is lower than the oxidation rate of SSB, the
respirometric response to XCB is slower and can
normally be identified quite easily as a tail to the
respirogram.
In a batch test an exponentially decreasing ‘tail’ can
frequently be observed in respirograms after the initial SS
peak (Figure 3.39). In Figure 3.42, this tailing starts after
approx. 0.75 hour. The wastewater concentration of XCB
can be assessed in a similar way to the above, using the
XCB
appropriate rO2,exo
(Kappeler and Gujer, 1992).
Simultaneously occurring oxidation processes such as
nitrification might interfere with and complicate the
identification of the respiration rate governed by
hydrolysis. In this case a nitrification inhibitor may be
used to facilitate the assessment of XCB (Spanjers and
1
2
3
4
5
6
Time (h)
Figure 3.42 Respiration rate (rO2) obtained according to Kappeler and
Gujer (1992) for estimation of XCB.
The example of Figure 3.43 is presented to stress that
all the methods mentioned above using oxygen
respiration are also applicable for anoxic respiration tests
using nitrate measurements (by lab analysis or using
nitrate probes). In this figure not a nitrate respirogram is
presented but, in fact, its integral, i.e. the change in nitrate
concentration. Backtracking through the different
periods with different nitrate utilization rates (the initial
period with SB and XCB biodegradation, followed by a
period with biodegradation of XCB only, followed by
endogenous nitrate respiration) to the Y-axis allows
direct assessment of the nitrate used for the different
processes and thus calculation of their concomitant
concentrations in the wastewater.
The calculation is to be done in a similar way as for
the oxygen respiration rate, with some modifications (the
respective ΔNOX of Figure 3.43 substitute the integrals
in this case):
SB (0) =
t
2.86 fin SB
rNOx,exo ( t ) dt
1 − YOHO,ax 0
XCB (0) =
t
2.86 fin XCB
rNOx,exo ( t ) dt
1 − YOHO,ax 0
Eq. 3.28
Eq. 3.29
168
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Where the COD equivalence term of 2.86 g COD g
NO3-N-1 is used. It should be taken into account that the
anoxic yield is somewhat lower than the aerobic yield:
whereas for YOHO a value of 0.66 g CODbiomass g
CODsubstrate-1 is often used, the anoxic yield is suggested
to be taken as 0.54 g CODbiomass g CODsubstrate-1 (Muller et
al., 2003).
3.4.5.5 Ammonium (SNHx)
25
20
NOx (mg N L-1 )
nitrifiers in the wastewater. However, it could be
imagined that a similar procedure as the one developed
for XOHO is applicable, i.e. evaluate the respiration rate
for nitrification rNit
O2, exo of the autotrophs present in the
wastewater and compare it to the respiration rate of a
culture with known autotrophic biomass concentration
XANO, e.g. after significant growth.
∆NOX for S B calculation
15
∆NO X for XCB calculation
10
5
0
0
1
2
3
4
5
6
Time (h)
Figure 3.43 A typical nitrate concentration profile for determination of SB
and XCB from a nitrate (i.e. anoxic) respiration experiment (Urbain et al.,
1998).
The wastewater ammonium concentration can be
assessed by using conventional analytical techniques.
However, respirometry also offers the possibility to
deduce SNHx from batch measurements in a similar way
to SB and XCB (provided the test is done with nitrifying
activated sludge). As follows from Table 3.2, the
nitrifiers yield coefficient YANO is needed to convert the
oxygen consumption for nitrification to nitrogen
concentration by dividing by (4.57 - YANO). However, the
value of SNHx is not very sensitive to YANO as YANO is
small compared to 4.57. Notice that ammonia is also used
for assimilation (i.e. about 12 % of the biomass weight as
VSS consists of nitrogen), which may require a
considerable amount of the nitrogen if a large amount of
COD is biodegraded (CODDegraded). The nitrogen that can
be nitrified can be approximated by:
N Nit = S NHx − i N,Bio · YOHO · COD Degraded
3.4.5.3 Heterotrophic biomass (XOHO)
Some wastewaters can contain significant concentrations
of heterotrophic biomass (Henze, 1992), so there is a
need to quantify this component. A batch test described
by Kappeler and Gujer (1992) and Wentzel et al. (1995)
assessed XOHO from the evolution of the respiration rate
of raw wastewater without adding a biomass seed. The
calculation requires YOHO and two parameters that can be
assessed from the same data: the maximum specific
growth rate μOHO and the decay coefficient bOHO.
Respirograms look like the one presented in Figure 3.40.
The procedure basically backtracks the amount of
heterotrophic biomass originally present in the
wastewater by comparing the original respiration rate
with the respiration rate after significant (hence, well
quantifiable) growth of XOHO.
3.4.5.4 Autotrophic (nitrifying) biomass (XANO)
So far, the authors are not aware of procedures in
which the autotrophic biomass concentration in
wastewater is assessed. This is probably due to the fact
that it is quite unlikely to find significant amounts of
Eq. 3.30
Where iN,Bio is the nitrogen content of newly formed
biomass. From this equation one can easily deduce the
original nitrogen concentration when CODDegraded, and the
stoichiometric parameters iN,Bio and YOHO are known.
Fitting a model in which carbon and nitrogen oxidation
are included to the respirometric data (i.e. the
optimisation method) will automatically take this
correction into account (Spanjers and Vanrolleghem,
1995).
3.4.5.6 Organic nitrogen fractions (XCB,N and SB,N)
Probably because the hydrolysis and ammonification
rates of organic nitrogen compounds are relatively fast,
little attention has been devoted so far to the
establishment of respirometric techniques for XCB,N and
SB,N quantification (these components have even been
abandoned in the subsequent Activated Sludge Models
No. 2 and 3). In batch tests, these compounds are
typically converted to SNHx before the SNHx that was
originally present in the wastewater has been removed by
nitrification. These fractions are thus encapsulated in the
RESPIROMETRY
determined ammonia concentration. Therefore, XCB,N
and SB,N are not directly observable in such tests.
Still, for some wastewaters the hydrolysis and
ammonification steps may be considerably slower and
quantification of the component concentrations may be
required. In such cases, one can imagine a procedure in
which the nitrification respiration rate rNit
is
O2, exo
monitored and interpreted in terms of hydrolysis and
ammonification, similar to the way the respiration
resulting from COD degradation is interpreted in terms
of the hydrolysis process. Subsequently, the amounts of
nitrogen containing substrates could be assessed by
taking the integral of rNit
O2, exo for the corresponding
fractions and dividing these by (4.57 ‒ YANO), the
stoichiometric coefficient corresponding to nitrification.
In case simultaneous COD removal is taking place,
correction should again be made for nitrogen assimilated
into the new heterotrophic biomass (Section 3.4.5.5).
3.5 BIOMASS CHARACTERIZATION
3.5.1 Volatile suspended solids
In biomass characterization tests, activity is usually
normalized by expressing a conversion rate per unit of
volatile suspended solids (VSS) in order to account for
the varying biomass concentrations. Note, however, that
VSS is not equivalent to biomass, that is: only a (usually
unknown) part of the VSS consists of active biomass, and
VSS is only an approximation of the concentration of
biomass. In fact, the challenge of the methods described
in this section is to assess the part of the VSS that
represents the biomass concentration. For further
information on how to assess the part of the VSS that is
active biomass the reader is referred to Ekama et al.
(1996), Still et al. (1996) and Lee et al. (2006).
3.5.2 Specific methanogenic activity (SMA)
169
biomass prior to its use as an inoculum for the start-up of
a new reactor and thus its potential as an inoculum for
that specific process (Sorensen and Ahring, 1993).
3.5.2.2 General
In SMA tests, the activity of the methanogens is
quantified. This is done by supplying a substrate that can
be converted directly into methane. This can be CO2 or
H2 gas, but a more common and a more practical
substrate is acetate. In literature this has become a
standard test as more and more researchers are using it
(Ersahin et al., 2014; Jeison and van Lier, 2007). The
conversion rate of acetate to methane, normalized to
biomass, gives information about the activity of the
methanogens in the biomass. It is usually expressed in g
COD g VSS-1 d-1.
To prevent acidification of the anaerobic digestion,
VFA production and VFA consumption need to be in
equilibrium (Section 3.4.1). The SMA therefore indicates
the maximum acetate production during the anaerobic
digestion that can be handled by the aceticlastic
methanogens without the pH dropping to values that
inhibit methanogenesis.
3.5.2.3 Test execution
In SMA tests (like BMP tests), the amount of inoculum
(required for anaerobic digestion) is usually measured as
volatile suspended solids (VSS), although volatile solids
(VS) is also possible (Section 3.4.1).
In an SMA test the ratio between the inoculum and
substrate is balanced such that the quantity of methane is
high enough to be measured accurately; the flow is to be
within the range of the gas flow meter and the
concentration of acetate is below inhibition levels. What
is often used is an acetate concentration of 2 g COD L-1
and the following substrate to inoculum ratio can be used:
VSSinoculum
CODsubstrate
3.5.2.1 Purpose
2=
The specific methanogenic activity (SMA) test involves
the assessment of the aceticlastic methanogenic activity
of a biomass. This is in contrast to the anaerobic biomass
activity tests where the overall activity of the biomass,
degrading a usually complex substrate, is measured; in
these tests the activity is limited by the rate of the slowest
degradation step, which is usually the hydrolysis so that
the hydrolytic activity of the biomass for that particulate
substrate is assessed. The SMA test may be used for
monitoring reactor performance or for characterizing
Both COD and VSS are given in masses supplied to
the test, expressed in grams, not concentrations.
•
Eq. 3.31
Stock solutions
- The stock solution needed is a phosphate buffer
which contains stock solution A (0.2 M
K2HPO4·3H2O: 45.65 g L-1) and stock solution B
(0.2 M NaH2PO4·2H2O: 31.20 g L-1).
170
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
-
The macronutrient solution contains per litre: 170
g NH4Cl, 8 g CaCl2·2H2O and 9 g MgSO4·7H2O.
The micronutrient solution contains per litre: 2 g
FeCl3·4H2O, 2 g CoCl2·6H2O, 0.5 g MnCl2·4H2O,
30 mg CuCl2·2H2O, 50 mg ZnCl2, 50 mg HBO3,
90
mg
(NH4)6Mo7O2·4H2O,
100
mg
Na2SeO3·5H2O, 50 mg NiCl2·6H2O, 1 g EDTA, 1
mL HCl 36%, 0.5 g resazurine, and 2 g yeast
extract.
• Inoculum
The inoculum is typical anaerobic sludge from a fullscale biogas plant. Measure total suspended solids (TSS)
and total volatile suspended solids (VSS) of the sludge.
• Substrates
As substrate for aceticlastic methanogenesis activity,
sodium acetate-trihydrate salt (NaC2H3O2·3H2O) (M =
136.02 g mol-1) is used. COD value: 0.4706 g COD per g
NaC2H3O2·3H2O. Typically, a sodium acetate solution
(substrate solution) with a concentration of 2.0 g COD
L-1 is prepared, which should include the following stock
solutions:
- Phosphate buffer: mix 30.5 mL of stock solution
A and 19.5 ml of stock solution A (in total 50 mL)
per litre substrate medium to obtain a 10 mM
phosphate buffer at pH = 7.
- Macronutrients: dose 6 mL per L substrate
medium.
- Micronutrients: dose 6 mL per L substrate
medium.
Check the pH and measure the COD concentration of
the substrate solution. For the blanks, prepare the same
solution but without substrate (medium solution).
• Preparation of the reaction vessels and test execution
Use triplicates for both the blank and sample, i.e. three
bottles as blanks (only inoculum and media solutions)
and three bottles for the sample (inoculum and substrate
solutions). As stated above, an inoculum to substrate ratio
of 2:1 (based on VSS) is normally used in the SMA test.
Choose the total volume of liquid that is suitable for the
reaction vessels. To perform the SMA test, as for the
BMP test, the bottles need to be incubated at a desired
temperature. The standard mesophilic temperature is 35
°C. The SMA test is continued until biogas production
ceases.
3.5.2.4 Data processing
The methane produced during an SMA test is measured
over time. In Figure 3.44, example data of an SMA test
are displayed. Notice that the test did not start at 0 N-mL.
This was because the head space of the reaction bottle
expanded during the warming up to 35 °C, creating a
small amount of gas flow. This gas flow in the first 10
minutes does therefore not represent the actual methane
flow. During the first two days a phase of low activity
can be observed: the lag phase. This is explained by the
fact that the methanogens at t = 0 were introduced into a
new matrix where the osmotic pressure, conductivity,
nutrient composition or substrate concentrations were
different from the matrix where methanogens were taken
from (digested sludge). This lag phase is normal for SMA
tests. After t = 4 days, the methane production has almost
come to a halt. This is most likely because of depletion
of the substrate. Prior to the start of an SMA test, the
expected methane production, based on the acetate COD,
can be calculated. For 1 g COD of acetate, theoretically
350 N-mL of methane can be produced. This can be
calculated as follows: 1 mole of acetate yields 1 mole of
CH4, hence 1 g Ac yields 1/59 mole of CH4. At standard
temperature and pressure (STP, that is: 273.15 K and
1013.25 mbar), 1 mole of gas is equivalent to 22.4 L,
hence 1 g Ac yields 22.4/59 = 0.380 L CH4. Since 1 g of
Ac represents a COD of 1.085 we get 0.380/1.085 =
0.350 L CH4 g COD-1.
180
y=35x-20
160
140
Methane (N-mL)
-
120
100
80
60
40
20
0
0
1
2
3
4
5
6
7
Time (d)
Figure 3.44 Example of an SMA test.
The data that needs to be extracted from this graph is
the maximum rate of production. This maximum rate of
production can be calculated from the slope of the
respirogram in Figure 3.44. The slope can be calculated
at any point, but the appropriate interval should be
RESPIROMETRY
171
chosen. In this case the interval between t = 2 and t = 4
(in red in Figure 3.44) is chosen because the production
rate is the highest here. The maximum production rate of
the sludge is 35 N-mL d-1. In this example 800 mg VSS
of inoculum was introduced and therefore the SMA is 43
N-mL g VSS-1 d-1. Note that the most common way of
expressing the amount of methane is in COD. One g
COD of methane occupies 350 N-mL of gas, at STP.
Therefore the SMA of this experiment is 43/350 = 0.123
g COD g VSS-1 d-1. This value falls within the typical
range of 0.1 and 0.2 g COD g VSS-1 d-1 that is commonly
found in conventional sludge digesters at WWTPs.
Sludge in a UASB-type reactor usually has a higher
SMA, between 0.2 and 0.4 g COD g VSS-1 d-1.
3.5.3 Specific aerobic and anoxic biomass
activity
Kristensen et al. (1992) summarized a set of laboratory
procedures that have been developed to assess the
activity of nitrifiers (AUR: ammonia utilization rate),
denitrifying biomass (NUR: nitrate utilization rate) and
heterotrophic biomass (OUR: oxygen utilization rate).
The three procedures and schematic data examples are
presented in Figure 3.45 and discussed below.
3.5.3.1 Maximum specific nitrification rate (AUR)
To assess the ammonia utilization rate (AUR),
concentrated biomass (e.g. taken from the return or
wastage line) and tap water are mixed in one litre
cylinders to reach a suspended solids concentration of 34 g VSS L-1. The activated sludge is kept in suspension
by aeration through diffusors, which also provide the
biomass with oxygen in a concentration range of 6-8 mg
O2 L-1. After reaching a stable condition (endogenous
respiration), ammonia is added to reach an initial NH4-N
concentration of 20 mg N L-1. Please note that
nitrification is an acidifying reaction and pH should be
monitored during the test to ensure that no pH effects are
occurring. The addition of alkalinity or installing a pH
control system with base addition can improve the quality
of the AUR determination.
• Recommendations
Especially in SMA tests it is important to pay attention to
the sensitivity of methanogens to temperature and
inhibition by oxygen. For further information on this and
some other recommendations the reader is referred to the
section on BMP tests.
The way to process the data of a blank group is
different from the BMP test, that is: in the SMA test the
data of the blank group is not subtracted from the data of
the sample group.
AUR
NUR
OUR
NO3+COD N2
NH4+O2
DO = 6-8 mg L-1
O2+COD+ATU
DO = 0 mg L-1
DO–probe
NH4
NO3
Time
OUR (mgO2 L-1)
NUR (mgNO3 L-1)
AUR (mgN L-1)
Air
Time
Time
Figure 3.45 Laboratory respirometric procedures and data examples for characterization of nitrifying biomass (AUR: ammonia utilization rate), denitrifying
biomass (NUR: nitrate utilization rate) and heterotrophic biomass (OUR: oxygen utilization rate) (Kristensen et al., 1992).
172
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Samples of 10 mL of activated sludge are then
withdrawn at intervals of 15-30 min for 3-4 h. The
samples are to be immediately filtered to stop the bioreactions, and the filtrates can be preserved by addition
of 0.1 mL of 4 M H2SO4 until the analysis. The samples
are subsequently analysed for ammonia nitrogen, and
nitrate plus nitrite nitrogen. Alternatively an ion-selective
ammonia probe, or otherwise an ion-selective or UVbased nitrate and nitrite sensor, can be used directly in the
aerated suspension to have more detailed time series and
potentially perform biokinetic modelling studies for the
nitrifiers.
The AUR (mg NH4-N L-1 h-1) is calculated from the
slope of the resulting nitrate plus the nitrite production
curve and as a control also from the ammonia utilization
curve. Indeed, ammonia uptake may also be affected by
endogenous heterotrophic activity due to decay and
ammonia release heterotrophic growth with concomitant
ammonia utilization. The produced oxidized nitrogen
forms are directly due to nitrification, of course on the
condition that oxygen is sufficiently high at all times to
prevent denitrification. The specific AUR (SAUR, mg
NH4-N g VSS-1 h-1) is obtained by dividing the
volumetric rate by the biomass concentration (g VSS L-1)
set at the beginning of the experiment to be able to
compare the nitrifying capacities with typical values.
nitrite utilization rate after all the ammonia has been
oxidized, one can calculate the activity of the nitriteoxidizing biomass. The AUR should now be obtained
from the ammonia profile, and not from the nitrate profile
since the latter is lagging behind the ammonia profile due
to the nitrite accumulation in the activated sludge. For
this type of analysis to work, it is important that the
ammonia-oxidizing capacity is significantly faster than
the nitrite-oxidizing capacity because sufficient nitrite
must accumulate (more than 2 mg NO2-N L-1 at ammonia
depletion is recommended).
An alternative method for determination of the
activities of the two biomass groups involved in
nitrification was developed by Surmacz-Gorska et al.,
(1996). It is based on a respirometric experiment to which
first sodium chlorate (NaClO3, 20 mM, i.e. 2.13 g L-1), an
inhibitor for the second step in nitrification, is added
when the DO has decreased by about 3 mg L-1, followed
by addition of ATU after DO declines with another 2 mg
L-1 to inhibit the first step of nitrification. A typical DO
concentration profile obtained in a closed bottle test is
given in Figure 3.46.
9
8
7
rO2,tot
from a respirometric experiment with the addition of
6
NaClO3
ammonia (Section 3.4.5.5). The maximum respiration
rate that can be attributed to nitrification (rNit
O,ex ) in mg O2
L-1 h-1, i.e. after subtraction of the endogenous
respiration, can be translated into an ammonia conversion
rate (mg NH4-N L-1 h-1) using an equation similar to the
one used to obtain the nitrifiable nitrogen (Section
3.4.5.5):
AUR =
rANO,O2
4.57 ‒ YANO
Eq. 3.32
Where YANO is the nitrifier yield coefficient, typically
0.24 mg COD mg N-1.
To separate the activities of the ammonia-oxidizing
and the nitrite-oxidizing biomass, it is possible to
perform two experiments, one with ammonia addition
and one with nitrite addition. The uptake of ammonia and
nitrite can be monitored in these separate experiments
and translated into the respective activities. Experiments
in which nitrite build-up occurs can also be used to
extract both activities separately, i.e. by determining the
DO (mg L-1)
Nitrification capacity can of course also be deduced
5
-
4
rO2,tot without NO2 oxidation
ATU
3
2
1
-
+
rO2,tot without NO2 and NH4 oxidation
0
0
1
2
3
4
5
6
Time (min)
7
8
9 10
Figure 3.46 A typical DO concentration profile recorded with the two-step
nitrification characterization procedure based on chlorate inhibition of
nitrite oxidation (Surmacz-Gorska et al., 1996). The slopes of the DO profile
are used to assess the respective rO2 (OURs).
The respective respiration rates, expressed in mg O2
L-1 h-1, are directly obtained from the three DO declines
and allow oxygen consumption to be calculated for the
two nitrification steps: rO2,NO2,exo (associated to rNOO,O2) is
calculated from the difference in the DO slope before and
after chlorate addition, whereas rO2,NH4,exo (associated to
rAOO,O2) is obtained from the difference in the DO slope
RESPIROMETRY
173
before and after ATU addition. With these respiration
rates, the activities of both nitrification biomasses (in mg
NH4-N L-1 h-1 and mg NO2-N L-1 h-1, respectively) can be
calculated as follows:
rNH4_NO2 =
rAOO,O2
3.43 ‒ YAOO
rNO2_NO3 =
rNOO,O2
1.14 ‒ YNOO
Eq. 3.33
Eq. 3.34
Where, YAOO and YNOO are the yield coefficients of
the two nitrification steps, typically 0.18 mg COD mg
NH4-N-1 and 0.06 mg COD mg NO2-N-1 respectively.
Specific activities are again obtained by dividing the
volumetric rates by the VSS concentration.
3.5.3.2 Maximum specific aerobic heterotrophic
respiration rate (OUR)
To determine the oxygen utilization rate related to
aerobic heterotrophic activity (OUR), biomass and tap
water are mixed to obtain a concentration of suspended
solids of 2-3 g VSS L-1 in a batch volume of one litre. An
experiment is then conducted with COD in excess.
Typically acetate is added in a concentration of typically
200 mg COD L-1, i.e. an S0/X0 ratio of about 1/10 to 1/20.
To obtain aerobic heterotrophic endogenous activity, an
alternative indicator of aerobic heterotrophic activity, no
COD is to be added. Nitrification must be inhibited by
the addition of allylthiourea (ATU, typically 5-10 mg
L-1). The biomass is continuously aerated to maintain a
DO concentration of 6-8 mg O2 L-1. In the procedure
proposed by Kristensen et al. (1992), the respiration rate
is measured by periodically pouring part of the batch into
a 300 mL BOD flask to measure the oxygen utilization
rate using an oxygen probe introduced into the flask (the
LSS principle). OUR can then be calculated from the
slope of the resulting DO decline. Note that here the total
respiration rate in the presence of a nitrification inhibitor
is used as the indicator of aerobic heterotrophic activity,
i.e. rO2,endo + rO2,exo. The alternative ways of obtaining the
respiration rate discussed in Sections 3.2 and 3.3 can of
course be applied to obtain OUR.
The specific oxygen utilization rate (SOUR), an often
used indicator of biomass activity, is calculated by
dividing the OUR by the VSS concentration in the batch
experiment.
An important element to consider is that the use of
acetate may not be ideal in all cases because some
biomasses may have adapted to a feast-or-famine regime
and store COD rather than use it directly for growth. The
corresponding respiration rate for storage may be quite
different than the respiration rate that is preferable for
heterotrophic activity assessment. If storage is detected,
an alternative COD source should be used for activity
assessment, e.g. the previously mentioned BOD
reference substrate glutamic acid.
3.5.3.3 Maximum specific denitrification rate (NUR)
The specific nitrate utilization rate (SNUR) for
denitrification can be assessed by using completely
mixed, closed atmosphere, two-litre batch reactors.
Concentrated biomass is collected and mixed with tap
water in the reactors to obtain a suspended solids
concentration of 3-4 g VSS L-1. After reaching a stable
condition (endogenous respiration), nitrate is added to
obtain an initial concentration of 20-30 mg N L-1. COD,
most often as acetate, is added in excess to obtain an
initial concentration of 100-150 mg COD L-1. For
determination
of
endogenous
NUR,
higher
concentrations of biomass are to be applied (to reduce the
experimentation time) and no COD is added. Since
denitrification is a process that increases pH, which is
amplified when acetate is the COD source (see pH
control results of NUR tests with acetate and glucose in
Figure 3.47; Petersen et al., 2002b), the pH should be
monitored and corrected for if necessary. Still, the
probability of negative pH impacts is much lower than
with AUR tests given the lower pH sensitivity of
heterotrophic biomass and the lower pH impact of the
denitrification.
Samples of 10 mL of activated sludge are withdrawn
at intervals of 15-30 min for 3-4 h. Samples are best
withdrawn under nitrogen gas addition in order to avoid
oxygen intrusion into the reactors. The samples are to be
pre-treated as mentioned above for the AUR
determination and analysed for nitrate plus nitrite
nitrogen. Alternatively, an ion-selective or UV-based
nitrate and nitrite sensor can be used in the reactor to have
more detailed time series and potentially perform
biokinetic modelling studies on the denitrification
process. An example of such data is given in Figure 3.47.
The NUR (mg NO3-N L-1 h-1) is calculated from the
slope of the resulting nitrate plus nitrite utilization curve.
The Specific NUR (SNUR, mg NO3-N g VSS-1 h-1) is
obtained by dividing the volumetric rate by the biomass
concentration (g VSS L-1) set at the beginning of the
experiment to be able to compare the denitrifying
capacities with typical values.
174
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
40
the specific conversion rates are lower under anoxic
conditions or only a fraction of the total biomass is
capable of respiring with nitrate. An often-used biomass
characteristic for this phenomenon is the so-called anoxic
reduction factor η that makes the ratio between both
activities on an electron-equivalent basis (hence the
factor 2.86):
A)
NO3-N (mg L-1)
30
20
10
η = 2.86·
rNO3 ,exo
rO2,exo
Eq. 3.35
0
40
60
80
120
Time (min)
20
B)
NO3-N (mg L-1)
15
10
5
0
0
20
40
60
80
100
Time (min)
Figure 3.47 Example of a nitrate utilization rate experiment with acetate
(A) and glucose (B) under pH-control with acid (A) or base addition (B)
(Petersen et al., 2002b).
Anoxic and aerobic heterotrophic activities are
closely related as they both reflect the capacity of
heterotrophic biomass to oxidize organic matter, with
either oxidized nitrogen or oxygen as the electron
acceptor. Figure 3.48 shows a typical comparison of
NUR and OUR respirograms obtained with the same
biomass and acetate addition (Sin and Vanrolleghem,
2004). The tailing-off occurring in both experiments after
return to endogenous respiration shows that the
aforementioned COD storage occurs both under anoxic
and aerobic conditions for this biomass sample.
When comparing respiration rates under anoxic and
aerobic conditions, one should be aware that the COD
conversion rates are, however, typically lower under
anoxic conditions. This is explained in two ways: either
The factor η is close to one (~0.85) for readily
biodegradable substrate (Ekama et al., 1996). For slowly
biodegradable substrate in the activated sludge model
No. 1 (ASM1), η ~0.33 (van Haandel et al., 1981) and in
ASM2, it is 0.66 (Clayton et al., 1991). The reason for
this difference has still not been understood.
Use of the ratio η assumes that the growth yield on
nitrate or on oxygen is the same. However, the growth
yield under anoxic conditions is typically lower than that
under aerobic conditions (Muller et al., 2003). This
means that for the same COD conversion rate more
electron acceptor will be consumed under anoxic
conditions (more COD must be burnt to achieve the same
biomass production). Therefore this means that, in
theory, the reduction factor may be above one if the COD
conversion rate is the same.
0.9
0.200
0.8
0.175
0.7
0.150
0.125
0.6
0.150
0.5
0.125
0.4
0.100
0.3
0.075
0.2
0.1
0.050
rNO3
rO2
rNO3 (mg NO3-N L-1 min -1)
20
rO2 (mg O2 L-1 min-1)
0
0.025
0.0
0.000
0
50
100
150
200
250
300
Time (min)
Figure 3.48 A comparison of aerobic (red) and anoxic (blue) respirograms
with acetate addition to sludge from WWTP Ossemeersen: after the
addition of 46.9 mg CODAc L-1 (rO2) and 38.9 mg CODAc L-1 (rNO3) (Sin and
Vanrolleghem, 2004).
RESPIROMETRY
175
References
Alberts, B., Johnson A., Lewis J., Raff, M., Roberts, K. and Walter
P. (2002) Molecular Biology of the Cell. 4th edition, Garland
Science, 1616 pp.
Angelidaki, I., Alves, M., Bolzonella, D., Borzacconi, L., Campos,
J.L., Guwy, A.J., Kalyuzhnyi, S., Jenicek, P., van Lier, J.B.,
2009. Defining the biomethane potential (BMP) of solid
organic wastes and energy crops : a proposed protocol for batch
assays. Water Sci Tech. 59(5): 927-34.
American Public Health Association (APHA), American Water
Works Association (AWWA), and Water Environment
Federation (WEF) (2012). Standard Methods for the
Examination of Water and Wastewater, 22nd Edition. New
York. ISBN 9780875530130.
Batstone, D.J., Keller, J., Angelidaki, I.., Kalyuzhnyi, S.,
Pavlostathis, S.G., Rozzi, A., Sanders, W., Siegrist, H. and
Vavilin, V. (2002) Anaerobic Digestion Model No. 1 (ADM1).
IWA Publishing, London.
Clayton, J.A., Ekama, G.A, Wentzel M.C. and Marais G.v.R (1991)
Denitrification kinetics in biological nitrogen and phosphorus
removal systems treating municipal wastewaters. Water Sci
Tech. 23(4/6-2): 1025-1035.
Coen, F., Petersen, B., Vanrolleghem, P.A., Vanderhaegen, B. and
Henze, M. (1998) Model-based characterization of hydraulic,
kinetic and influent properties of an industrial WWTP. Water
Sci Tech. 37(12): 317-326.
Copp, J.B., Spanjers, H., Vanrolleghem, P.A. (Eds.), 2002.
Respirometry in control of the activated sludge process:
Benchmarking control strategies. Scientific and Technical
Report No. 11, IWA Publishing, London, UK.
Eastman, J.A. and Ferguson, J.F., 1981. Solubilization of particulate
organic carbon during the acid phase of anaerobic digestion.
Water Poll. Control Fed. 53: 352-366.
Ekama, G.A., Dold, P.L. and Marais, G.v.R. (1986) Procedures for
determining influent COD fractions and the maximum specific
growth rate of heterotrophs in activated sludge systems. Water
Sci Tech. 18: 91-114.
Ekama, G.A., Wentzel, M.C., Casey, T.G. and Marais, G.v.R (1996)
Filamentous organism bulking in nutrient removal activated
sludge systems. Paper 3: Stimulation of the selector effect
under anoxic conditions. Water SA, 22(2):119-126.
Ersahin, M.E., Ozgun, H., Tao, Y., van Lier, J.B., 2014.
Applicability of dynamic membrane technology in anaerobic
membrane bioreactors. Water Res. 48: 420-9.
Gaps, D., Pratt, S., Yuan, Z., and Keller, J. (2003) Online titrimetric
and off-gas analysis for examining nitrification processes in
wastewater treatment. Water Res. 37: 2678-2690.
Gernaey, K., Petersen, B., Ottoy, J.P. and Vanrolleghem, P.A.
(1999) Biosensing activated sludge. WQI, May/June 1999, 1621.
Henze, M. Grady Jr., C.P.L., Gujer, W., Marais G.v.R. and Matsuo,
T. (1987). Activated sludge model No. 1. Scientific and
Technical Report No. 1, IAWPRC, London.
Henze, M. (1992) Characterization of wastewater for modelling of
activated sludge processes. Water Sci Tech. 25 (6): 1-15.
Henze, M., Harremoes, P., Jansen, J. and Arvin, E. (1995).
Wastewater Treatment Biological and Chemical Processes.
Berlin, Springer-Verlag, pp. 383.
Jeison, D. and van Lier, J.B., (2007). Thermophilic treatment of
acidified and partially acidified wastewater using an anaerobic
submerged MBR: Factors affecting long-term operational flux.
Water Res. 41: 3868-79.
Kappeler, J. and Gujer, W. (1992) Estimation of kinetic parameters
of heterotrophic biomass under aerobic conditions and
characterization of wastewater for activated sludge modelling.
Water Sci Tech. 25(6): 125-139.
Kong, Z., Vanrolleghem, P.A. and Verstraete, W. (1994)
Automated respiration inhibition kinetics analysis (ARIKA)
with a respirographic biosensor. Water Sci Tech. 30(4): 275284.
Kong, Z., Vanrolleghem, P.A., Willems, P. and Verstraete, W.
(1996) Simultaneous determination of inhibition kinetics of
carbon oxidation and nitrification with a respirometer. Water
Res., 30: 825-836.
Kristensen, G.H, Jorgensen, P.E. and Henze, M. (1992)
Characterization of functional groups and substrate in activated
sludge and wastewater by AUR, NUR and OUR. Water Sci
Tech. 25(6): 43-57.
Lasaridi, K.E. and Stentiford, E.I. (1998) A simple respirometric
technique for assessing compost stability. Water Res. 32(12):
3717-23.
Lee, B.J., Wentzel, M.C. and Ekama, G.A. (2006) Measurement and
modelling of ordinary heterotrophic organism active biomass
concentration in anoxic/aerobic activated sludge mixed liquor.
Water Sci Tech. 54(1):1-10.
Muller, A., Wentzel, M.C., Ekama, G.A. and Loewenthal, R.E.
(2003) Heterotrophic anoxic yield in anoxic aerobic activated
sludge systems treating municipal wastewater. Water Res. 37:
2435-2441.
Nelson, D.L., and Cox, M.M. (2008) Lehninger Principles of
Biochemistry. 5th Edition, Palgrave Macmillan, ISBN: 978-0716-77108-1, 1100 pp.McCarthy, E.L. (1934). Mariotte’s
bottle. Science. 80, 100.
Petersen, B., Gernaey, K., Henze, M. and Vanrolleghem, P.A.
(2002a) Evaluation of an ASM1 model calibration procedure
on a municipal-industrial wastewater treatment plant. J.
Hydroinformatics, 4, 15-38.
Petersen, B., Gernaey, K. and Vanrolleghem, P.A. (2002b) Anoxic
activated sludge monitoring with combined nitrate and
titrimetric measurements. Water Sci Tech. 45(4-5): 181-190.
Petersen, B., Gernaey, K., Henze, M. and Vanrolleghem, P.A.
(2003) Calibration of activated sludge models: A critical review
of experimental designs. In: Biotechnology for the
Environment: Wastewater Treatment and Modeling, Waste Gas
Handling. Eds. Agathos S.N. and Reineke W., Kluwer
Academic Publishers, Dordrecht, The Netherlands. 101-186.
Pratt, S., Yuan, Z., Gapes, D., Dorigo, M., Zeng, R.J. and Keller, J.
(2003) Development of a Novel Titration and Off-Gas Analysis
(TOGA) Sensor for Study of Biological Processes in
Wastewater Treatment Systems. Biotechnol. Bioeng. 81: 482495.
Rezouga, F., Hamdi, M. and Sperandio, M. (2009) Variability of
kinetic parameters due to biomass acclimation: Case of paranitrophenol biodegradation. Bioresource Technology. 06/2009;
100(21): 5021-9.
Rieger, L., Langergraber, G., Kaelin, D., Siegrist H. and
Vanrolleghem P.A. (2008) Long-term evaluation of a spectral
sensor for nitrite and nitrate. Water Sci Tech. 57(10): 15631569.
Sin, G. and Vanrolleghem, P.A. (2004) A nitrate biosensor based
methodology for monitoring anoxic activated sludge activity.
Water Sci Tech. 50(11): 125-133.
Sorensen, A.H. and Ahring, B.K. 1993. Measurements of the
specific methanogenic activity of anaerobic biomass. Appl.
Microbiol. Biotechnol. 40: 427-431.
Spanjers, H. and Vanrolleghem, P.A. (1995a) Application of a
hybrid respirometric technique to the characterization of an
industrial wastewater. Proceedings 50th Purdue Industrial
Waste Conference, 611-618.
Still, D.A., Ekama, G.A., Wentzel, M.C., Casey, T.G. and Marais,
G.v.R (1996) Filamentous organism bulking in nutrient
removal activated sludge systems. Paper 2: Stimulation of the
176
selector effect under aerobic conditions. Water SA, 22(2): 97118.
Spanjers, H. and Vanrolleghem, P.A. (1995b) Respirometry as a
tool for rapid characterization of wastewater and activated
sludge. Water Sci Tech. 31(2): 105-114.
Spanjers, H., Vanrolleghem, P.A., Olsson, G., Dold, P.L., 1998.
Respirometry in control of the activated sludge process:
Principles. Scientific and Technical Report No. 7, IAWQ,
London, UK.
Surmacz-Gorska, J., Gernaey, K., Demuynck, C., Vanrolleghem,
P.A. and Verstraete, W. (1996) Nitrification monitoring in
activated sludge by oxygen uptake rate (OUR) measurements.
Water Res., 30:1228-1236.
Theodorou, M.K., Williams, B.A., Dhanoa, M.S., Mcallan, A.B.,
France, J., 1994. A simple gas production method using a
pressure transducer to determine the fermentation kinetics of
ruminant feeds. Anim. Feed Sci. Technol. 48: 185–197.
Urbain, V., Naidoo, V., Ginestet, P. and Buckley, C.A. (1998)
Characterization of wastewater biodegradable organic fraction:
accuracy of the nitrate utilisation rate test. In Proceedings of the
Water Environmental Federation 71st Annual Conference and
Exposition, October 3-7, Orlando, Florida (USA), 247–255.
van Haandel, A.C., Ekama, G.A. and Marais, G.v.R (1981) The
activated sludge process Part 3 - Single sludge denitrification.
Water Res., 15: 1135-1152.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Vanrolleghem, P.A., Kong, Z., Rombouts, G. and Verstraete, W.
(1994). An on-line respirographic biosensor for the
characterization of load and toxicity of wastewaters. J. Chem.
Technol. Biotechnol., 59: 321-333.
Vanrolleghem, P.A., Spanjers, H., Petersen, B., Ginestet, P. and
Takacs, I. (1999) Estimating (combinations of) Activated
Sludge Model No.1 parameters and components by
respirometry. Water Sci Tech. 39(1): 195-214.
Volskay, V. and Grady, C. (1990) Respiration inhibition kinetic
analysis. Water Res., 24, 863-874.
Weijers, S.R. (2000) Modelling, Identification and Control of
Activated Sludge Plants for Nitrogen Removal. PhD thesis.
Eindhoven University of Technology.
Wanner, O., Kappeler, J. and Gujer, W. (1992) Calibration of an
activated sludge model based on human expertise and on a
mathematical optimization technique - A comparison. Water
Sci Tech. 25(6): 141-148.
Wentzel, M.C., Mbewe, A. and Ekama, G.A. (1995) Batch test for
measurement of readily biodegradable COD and active
organism concentrations in municipal waste waters. Water SA,
21, 117-124.
Zhang, X., Hu, J., Spanjers, H., van Lier, J.B., (2014). Performance
of inorganic coagulants in treatment of backwash waters from
a brackish aquaculture recirculation system and digestibility of
salty sludge. Aquac. Eng. 61: 9–16.
4
OFF-GAS EMISSION TESTS
Authors:
Reviewers:
Kartik Chandran
Eveline I.P. Volcke
Mark C.M. van Loosdrecht
Peter A. Vanrolleghem
Sylvie Gillot
4.1 INTRODUCTION
(WWTPs), and to develop a methodology to collect fullscale plant data in many parts of the world. In this
chapter, the approaches applied in several such studies
are described. One of the original protocols developed
by researchers in the US, which was reviewed and
endorsed by the United States Environmental Protection
Agency (USEPA), is presented in detail, as well as
methods from other studies elsewhere. It is expected and
approaches presented in this chapter will enable off-gas
and greenhouse gas emissions from WWTPs to be
monitored and quantified using standardized
experimental approaches. Because plant operation has a
key role in off-gas and greenhouse gas emissions, the
results of such monitoring and quantifying efforts could
ultimately lead to WWTPs not just improving their
aqueous discharges, but at the same time minimizing
their gaseous emission. This chapter focuses on the
emissions from wastewater treatment by activated sludge
and similar processes. Methane gas measurements are
well-established as a standard method, whereas an
overview of all nitrous oxide measurement methods is
given by Rapson et al. (2014).
Wastewater treatment technology and biological nutrient
removal (BNR) processes are used mainly to achieve
improved water quality. However, BNR processes are a
potential contributor to nitrous oxide (N2O) and methane
(CH4) emissions to the atmosphere, depending on the
plant configuration and operating conditions
(Kampschreur, 2009). For measuring the carbon
footprint of wastewater treatment and BNR processes,
this flux of greenhouse gases (GHGs) to the atmosphere
should be measured. GHG emissions can be further
enhanced by increasing attention on energy neutral
treatment plants based on methane generation by
anaerobic processes. The release of methane, and
especially of nitrous oxides, is of concern because the
greenhouse impact of N2O on a mass equivalent basis is
about 300 times that of carbon dioxide (CO2) and 34
times of CH4 (IPCC, 2013).
During the past few decades, there have been ever
increasing attempts to characterize off-gas (and GHGs)
emissions from full-scale wastewater treatment plants
© 2016 Kartik Chandran et al. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen. C.M. Lopez-Vazquez and D. Brdjanovic. ISBN:
9781780404745 (Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
178
Off-gas measurements of carbon dioxide and oxygen
(O2) are a good tool for evaluating plant performance.
The CO2 and O2 content in off-gas can be measured with
similar strategies described here for GHGs. A
combination of off-gas measurements and titration
measurements forms an excellent way to monitor
biological processes (Pratt et al., 2003; Gapes et al.,
2003). Oxygen depletion in the off-gas is a measure of
the respiration rates and can be seen as a (full-plant)
respirometric approach, similar to those described in
Chapter 3. Carbon dioxide production is also related to
the aerobic and anoxic respiration processes. However it
is more difficult to interpret since the change (usually a
net decrease) in alkalinity also significantly influences
the carbon dioxide content in the off-gas (Hellinga et al.,
1996). Measuring the oxygen and carbon dioxide
conversion could be of great value for better process
monitoring and data handling (see Chapter 5), since they
allow a better evaluation of mass balances over the
wastewater treatment process (Takacs et al., 2007; Puig
et al., 2008).
4.2 SELECTING THE SAMPLING STRATEGY
A major consideration for off-gas measurements from
WWTPS is a representative sampling procedure. Given
the biological nature of N2O production and link to
processes such as nitrification and denitrification, spatial
and temporal variability in the measured emission fluxes
of these gases is expected. A correlation with wastewater
constituents such as ammonia, nitrite and nitrate could
also be expected and has been observed in nearly all the
measurement campaigns conducted to date. Since
emissions from full-scale WWTPs of especially N2O
have proven to be highly dynamic, a measurement
program has to be decided upon in a case-by-case
evaluation (Daelman et al., 2013). The sampling
strategies employed during sampling campaigns must
adequately capture such variability. Some factors that
influence the sampling strategies are described below.
4.2.1 Plant performance
One of the most significant results of past studies has
been the implication that N2O emissions from WWTPs
are linked to overall or local accumulation of nitrogen
species such as ammonium (NH4+) or nitrite (NO2-).
There is especially a need for characterization of
emissions from plants that do not show a stable
nitrification or denitrification performance. Such
characterization is important given that many plants are
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
not designed or operated in a BNR mode. Under certain
conditions (such as high temperatures during the summer
when these highly-loaded plants exhibit incomplete
nitrification), these plants can emit more N2O in
comparison to well-designed and operated BNR plants.
In contrast to N2O, CH4 emissions are likely to occur due
to lack of control in an activated sludge process. For
instance, poor aeration or dissolved oxygen (DO) control
in the activated sludge process may lead to anaerobic
zones and associated CH4 production. Besides, CH4 that
was formed in the sewer system and enters the plant with
the influent is likely to be stripped in the activated sludge
tank, even though it could also be aerobically oxidized.
CH4 stripping in the activated sludge tank could be
minimized while its biological conversion could be
promoted through adequate process design and control
(Daelman et al., 2014)
4.2.2 Seasonal variations in emissions
The link between temperature and emissions of N2O or
CH4 is not straightforward and needs to be considered in
combination with the plant design. For plants operating
at a high sludge retention time (SRT) and sustaining
complete NH3 removal all year round, higher emissions
can be expected during the summer period, owing to
overall higher microbial activity at higher temperatures.
Especially for such plants, the off-gas emission may shift
more downstream in plug-flow installations at lower
temperatures. The slower nitrification occurs more
downstream at these lower temperatures. For plants that
cannot fully nitrify at low temperatures and foster nonlimiting NH3 concentrations at lower temperatures, offgas emission can actually be higher at lower
temperatures. In contrast, as these plants can achieve full
nitrification during the higher temperature, their
emissions could be lower. Seasonally nitrifying treatment
plants could have higher emission as well. Such plants
can ‘wash-out’ nitrifying bacteria at low temperatures
and consequently have extremely low off-gas emission.
As nitrification is supported during higher temperatures
but often not sufficiently to achieve the required effluent
NH3 concentration, off-gas emission could be higher
during the summer. For planning gas measurement
locations the measurement of nitrite might be helpful
since there is a general, but not unique, correlation
between the presence of nitrite in the liquid phase and
N2O formation. CH4 emissions from aeration tanks result
largely from processes in sewer and sludge processing
facilities, and therefore the link between temperature and
measured emissions is much less obvious than for N2O
emissions.
OFF-GAS EMISSIONS TESTS
4.2.3 Sampling objective
A range of different sampling strategies for the
quantification of N2O emissions has been reported in
literature, such as 24 h online sampling, 1-week online
sampling, long-term weekly grab sampling and taking a
single grab sample. The difference in the sampling
methods may partly explain the large variability in N2O
emissions reported, aside from the differences in the real
emission, which is related to the plant performance and
shows both diurnal and seasonal variations.
The optimal sampling strategy depends on the
objective of the sampling campaign. Daelman et al.
(2013) applied several N2O sampling strategies applied
in literature to the extensive dataset of a long-term online
monitoring campaign. They showed that a reliable
determination of the actual average N2O emission
requires long-term sampling, be it online or grab
sampling, covering the entire temperature range that can
possibly be encountered. Short-term sampling
unavoidably ignores the long-term variation that is
present. For long-term grab sampling, nighttime and
weekend samples contribute significantly to a more
accurate estimate. High-frequency (online) sampling is
indispensable to identify correlations between the
emission and the process variables that induce the
emissions and in this way gain insight into the underlying
N2O formation mechanisms. A method to obtain the
number of grab samples or online sampling periods
required to obtain a sufficiently precise estimation of the
emission was presented by Daelman et al., (2013),
serving as a guideline for sampling campaigns, to balance
cost and precision.
4.3 PLANT ASSESSMENT AND DATA
COLLECTION
4.3.1. Preparation of a sampling campaign
A sampling program (or campaign) is primarily carried
out to (i) collect routine operating data concerning the
overall WWTP performance, (ii) acquire information that
can be used to document the performance of a given
treatment operation or process, (iii) collect data that can
be used to implement proposed new programs for
sewerage and WWTP, and (iv) obtain records needed for
reporting regulatory compliance. In general this
information is:
179
a.
b.
c.
d.
e.
Influent characterization measurements.
Mass balance measurements.
Activated sludge characterization measurements.
Effluent characterization.
All other necessary information.
A well-prepared sampling plan should not only give
the type (e.g. chemical oxygen demand: COD, total
Kjeldahl nitrogen: TKN, total phosphorus: TP) and
location (e.g. influent, effluent or aeration tank) of the
sampling point, but it should also indicate what the
sampling frequency is and how the sample should be
taken, handled, stored, and processed in the laboratory
for analytical measurements. The goal of the campaign
planning is to communicate all this information to the
staff responsible for acquiring the samples (operators or
plant technical personnel) and the laboratory personnel
analysing the samples. Tables should be developed
indicating all of this information, preferably in a single
table overview. Also, a dedicated process flow diagram
should be developed indicating precisely all the
measurement points related to the planning.
While designing the sampling campaign one must
address at least the following practical questions:
1. WHY: What is the purpose of taking the samples?
2. WHAT: Which parameters are to be sampled?
3. WHERE: Where are the sampling locations?
4. WHEN: When are the samples taken?
5. HOW OFTEN: What is the frequency for each
sampled parameter?
6. HOW: How will the sampling be executed?
7. BY WHOM: Who will take the samples and who will
analyse them?
8. EQUIPMENT: What will be used for collecting,
storing and analysing the samples?
The six criteria for quality data are:
Collecting representative samples.
Formulating the objectives of the sampling program.
Proper handling and preservation of the samples.
Proper chain-of-custody and sample identification
procedures.
5. Field quality assurance.
6. Proper analysis.
1.
2.
3.
4.
The sample, in order to serve its purpose, has to be
representative, reproducible, defendable and useful. It is
also important to realize that proper understanding of the
sampling procedures is crucial for the design and success
of the sampling program. This should not only be realized
180
by the designer but also by all the others involved in the
project. Therefore, it is advised to organize a
preoperational meeting to communicate the measurement
planning to all the personnel involved. During this
meeting, it should be clearly stated what the goal of the
project is, what will be done with the specific
measurements and also the importance of acquiring
accurate information. It is possible that it will be
necessary to give (short practical) training in (grab)
sampling techniques.
To ensure that the measurement planning is properly
understood and executed, the measurement plan should
relate as much as possible to the practical routines of the
plant and laboratory personnel. Where possible,
references should be made to standard (practical)
routines. For example, measurement locations on the
process flow diagram should be indicated in the same
way as used by the operators. Often the designer has his
own preferred way of indicating the flows (e.g. in a flow
diagram using the logical numbering Q1 to Qn to indicate
flows in the mass balance), to communicate this
information effectively, and the practical names of the
process units and flows should also be indicated in all the
diagrams and tables.
Also, for analytical measurements the designer can
have different names that do not always correspond with
the names and practical methods applied in the
laboratory. Therefore the measurement plan should also
be checked with the laboratory personnel and the applied
measurement routines (analytical methods and
equipment).
For the purpose of a campaign, samples should
preferably be taken in dry weather conditions unless
otherwise specified. In practice it is not always clear if
and when these conditions are met. Typically it is advised
to plan a day of sampling after three days without (major)
rain. This means that the measurement planning can
never be planned on an exact date. However, it should
incorporate a period of several days in which the
planning can be executed, depending on the weather
conditions.
Some of the samples can be done automatically with
24-h continuous (or quasi-continuous) flow proportional
composite measurement devices. These automatic
samplers collect samples over a 24-h period and have to
be started 24 h ahead of time to provide a collected
sample on the desired sampling day. It should be born in
mind that operators have to be notified a day in advance
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
to be able to collect the samples on time. All the other
samples taken on the day of sampling will mostly be spot
samples (also called grab, random, catch, individual,
instantaneous, and snap samples). Because the sampling
program contains multiple measurement points and for
each measurement point, multiple parameters have to be
measured, the amount of containers required to store the
samples will rapidly increase; e.g. the amount of samples
needed for a relatively small campaign can easily exceed
100. Also, it should be realized that for each sampling
point the sampled volume should allow analytical
measurements, washing of filters and vessels, possible
spills and possibly duplicate or triple measurements in
the laboratory. Also, the laboratory should be prepared to
handle this large amount of samples in one (or more)
days. Some of the samples can easily be stored to be
measured at a different time. However, for other
measurements it is critical that the samples are pretreated (e.g. filtered) before they can be stored. Again,
other samples cannot be stored at all (e.g. as the result of
biological conversions which occur in the sample vessel).
Also, this should be planned in cooperation with the
laboratory personnel to avoid a possible work overload
and possible inaccurate measurement results.
4.3.2 Sample identification and data sheet
As was pointed out previously, in the sample
measurement plan many samples may need to be
analysed for one sampling day. Clearly this requires a
good administration and sample identification (sample
ID). A list of the samples collected should be kept. The
samples should be listed in one overview and include the
information below. In addition, the individual sample
containers should be labelled accordingly.
The container label requirements are:
a. Sample site location.
b. Container number if more than one
container/sample.
c. Name of sample collector.
d. Facility name/location.
e. Date and time of collection.
f. Identified as a grab or composite.
g. Test parameters - list analysis to be done.
h. Preservative used.
Proper administration of the sampling program is
necessary for quality assurance. The following checklist
can be used for constructing custody forms supporting
the measurement campaign. Often, forms of this kind will
already be available and used for the regular sampling
OFF-GAS EMISSIONS TESTS
181
program. It can be helpful to find out if these forms and
also the forms in the measurement campaign are
available instead of constructing new ones.
4.3.4 Practical advice for analytical
measurements
a. Project number: assign a number to the sampling
episode that can be used on the bottles to track the
samples.
b. Sampler: the sampler should print and/or sign their
name.
c. Project name: the project name, and if necessary, the
project address.
d. Project contact name: the name of the person who is
in charge of the project.
e. Project telephone number: the phone number where
the project contact person can be contacted.
f. Sample date: list the dates for grab and composite
samples.
g. Sample time: list the sample time for each sample.
h. Sample type: composite (collected over a 24-h
period) or grab (taken as grab samples).
i. Station location: the site where the sample was taken.
j. Number of containers: list the number and type of
containers for each sampling event.
k. Analysis requested: list the analysis needed that is
compatible with reference to official methods.
l. Remarks: note any special requirements for the
samples for the laboratory.
m. Split samples: when a sample is split into different
proportions for analysis, the receiving party can
accept or reject the samples and they must sign this
box on the custody form.
Some typical re-occurring errors in the case studies in the
sample procedure or analytical measurements have been
documented and are listed below:
• The method used for analytical measurements of
suspended solids is different for concentrated and
diluted activated sludge samples; if the concentration
is above a certain range, then filtration, drying and
weighing the solids will no longer be possible. In this
case the dry mass method is often used (evaporation
of the sample). This includes the soluble salts, which
especially for lower concentrated sludge samples,
will not give representative results for total
suspended solids. This should especially be taken into
account when measuring activated sludge, thickened
sludge and dirty water flows. For the more
concentrated sludge flows (> 20%), dry mass analysis
can be used.
• When measuring the filtrate of the sample containing
suspended solids, the filter paper should not be
washed other than with the sample itself and not with
demineralized water, which will cause dilution of the
measured filtrate.
• When handling samples containing suspended solids,
it is important to take well stirred homogeneous
samples when taking the samples or pouring a sample
into another vessel. It should be noted that TSS
measurements are essential for the model design
study and at the same time are very difficult to sample
accurately. It is advised to instruct all personnel on
how to take reliable TSS samples.
• When carrying out the activated sludge
characterization, all the samples should be taken from
the same vessel as it is important to know the relative
concentration of TSS (total suspended solids), VSS
(volatile suspended solids), COD, TKN and TP in the
sample. This is not usually a standard procedure and
therefore all the personnel, especially those working
in a laboratory responsible for analysing, should be
well instructed in this matter.
• A common known problem is the destruction of
suspended solids, which is necessary to analytically
measure TP (and also COD and TKN). When the
destruction procedure is not long enough, phosphorus
measurements will be underestimated. Especially
when there are chemical precipitates (e.g. as a result
of dosing of iron-salts) including phosphorus in the
sample, destruction can be difficult.
4.3.3 Factors that can limit the validity of
the results
When evaluating the results of the sampling program,
there are several factors that can limit the validity of the
results, such as:
a. Missing values.
b. Sampling frequencies that change over the recorded
period of time.
c. Multiple observations within the sampling period.
d. Uncertainty in the sample preservation and
measurement procedures.
e. Censoring the measurement signals.
f. Small sample size.
g. Improper data handling.
h. Equipment inaccuracies.
It is always useful to check the sampling procedure
and the measurement results against the above criteria.
182
•
•
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Measurements of nitrate in samples containing
activated sludge (denitrifying heterotrophic bacteria)
should be done quickly before denitrification occurs.
Measurement of volatile fatty acids (VFAs) in
samples containing activated sludge (denitrifying
heterotrophic bacteria) is practically impossible for
the same reason; VFAs are almost immediately
oxidized by living organisms. These types of samples
are highly unreliable and should not be used.
Measurement of volatile fatty acids in the influent is
possible; however, it should be immediately filtrated
after sampling and stored under prepared conditions.
4.3.5 General methodology for sampling
In general, there is a standard approach developed
regarding the design of a water quality monitoring
system; the approach for a sampling program at a WWTP
can be derived from this that consists of six distinctive
steps:
Step 1: Evaluation of the existing information
• Wastewater collection and treatment.
Domestic wastewater.
Industrial effluents.
Combined sewer overflows and pumping
stations.
Wastewater treatment plants.
• Lab-scale and pilot plant investigations.
Wastewater treatment plants.
Wastewater collection networks.
Step 2: Evaluation of the information expectations
• Water quality goals and objectives.
• Water quality problems and issues.
• Management goals and strategy.
• Monitoring role in management.
• Monitoring goals (as statistical hypotheses).
Step 3: Establishment of statistical design criteria
• Statistically characterise the population to be
sampled.
Variation in quality.
Seasonal impacts.
Correlation (independence).
Applicable probability distributions.
Out of the many statistical tests select the most
appropriate.
Step 4: Design the monitoring network
• Where to sample (from the monitoring role in
management).
• What to measure (from the water quality goals and
problems).
• When to sample (from specific circumstances).
•
How frequently to sample (from the needs of the
statistical tests).
Step 5: Develop the operating plans and procedures
• Sampling routes and points.
• Field sampling and analysis procedure.
• Sample preservation and transportation.
• Laboratory analysis procedures.
• Quality control procedures.
• Data management and retrieval hardware and
database management systems.
• Data analysis software.
Step 6: Develop reporting procedures
• Type of format of the reports.
• Frequency of report publication.
• Distribution of reports (information).
• Evaluation of the report’s ability to meet the initial
information expectations.
Based on the above, a general wastewater and sludge
sampling program can be carried out following the steps
listed below:
• Definition of the purpose of the sampling.
• Determination of the type, scope and required
accuracy of the analyses to be carried out.
• Definition of the character of the samples to be
collected.
• Selection of the localities and sources to be sampled,
and of the sampling points at these localities.
• Determination of the hydraulic and other parameters
relating to the subject of the sampling program.
• Consideration of the occupational safety and hygiene
of those collecting the samples.
• Preparation of an optimal sampling programme.
• Selection of the sampling and measuring equipment
suited to the sources to be sampled, and
determination of its state.
• Selection of the most suitable sampling technique in
line with a fixed programme and selected equipment
for the given source/site, including preparations,
subsidiary measurements and observations. When
implementing a sampling programme the techniques
used should comply with a number of general
requirements such as reliability, economy,
repeatability and conservation.
• Selection of appropriate procedures, sample handling
equipment and tools, transportation from site to
laboratory, storage after delivery.
• Consideration of the most suitable methods of
analysis on site and in the laboratory, the quickest
possible interpretation of the analyses including
reliability checks, and possible repetition of the
OFF-GAS EMISSIONS TESTS
•
•
•
•
sampling, and other factors likely to influence the
accuracy or representativeness of the analyses.
Use of feedback whenever possible to modify a
programme to an optimal sampling programme.
Establishment of the conditions necessary for the
immediate use of the results and for their storage as
primary sources of information in the future, whether
for short- or long-term use.
Selection of an appropriate system for data
management (i.e. handling, processing, transfer and
manipulation).
Selection of the method of documentation to be used
throughout the programme.
Logistically, performing the measurement campaign
in the design study is a complicated task and at the same
time it is the most important and critical step. Therefore
it is advised to carefully prepare the measurement
campaign in all possible aspects. It is also advised to
accompany the first rounds of the measurements together
with all the personnel involved to make sure that all the
instructions in the measurement plan are correct and as
intended, and properly understood by all the personnel.
The analytical results of a sample are only as accurate as
the quality of the sample taken. If the technique for
collecting samples is poor, then no matter how accurate
your lab procedures are, the results will be poor. By
sampling according to set procedures, one reduces the
chance of error and increases the accuracy of the sample
results. It should be born in mind that for a lot of samples
the measurement campaign is a one-off opportunity;
redoing samples at a different stage will often not lead to
satisfactory results because the activated sludge plant is
a dynamic system that is continuously changing over
time. With this in mind, for the study the samples should
be collected simultaneously in such a way that their
relative information can be used for the plant assessment.
Therefore, individual grab samples will often not be
representative, meaning that (missing or incorrect)
measurements often cannot be redone. In conclusion,
data collection through sampling at WWTPs is a critical
activity that needs a systematic and professional
approach and the design of a sampling program is often
a task for an experienced specialist.
183
teams need to gather process-operating data during
meetings with plant operators (and process engineers)
prior to the sampling campaign. The following
background information is typically collected from the
evaluation site:
• An overall plant description which includes general
information related to the plant configuration, liquid
and solids process flow diagrams, design criteria,
major technological process equipment from the
plant’s design documentation and/or operation and
maintenance manuals.
• Process units and layouts which entail information on
anaerobic/aerobic/anoxic zone configuration, zone
volumes, operating set points, basins in service,
aeration flow and distribution, recycle streams and
flow rates (if applicable), and
• Plant operating data, ideally providing a summary of
a minimum of 3 months’ of plant data on relevant
treatment process(es) to allow for characterization of
the influent and effluent, and target and actual
operating set points for key operational parameters
(e.g. DO, SRT). The operating data is needed to
check that the plant was not in an upset condition
before and during the period of investigation
(sampling).
The collected data and information can be used to
make analyses by conventional techniques such as the
development of solids and nitrogen mass balances as well
as through the use of process modeling. Details of modelbased evaluation are not presented in this chapter and can
be found elsewhere (Melcer, 1999; Puig et al., 2008;
Rieger et al., 2012).
4.3.6 Sampling in the framework of the offgas measurements
In order to collect the necessary data for the
assessment it is necessary to determine the influent flow,
and organic matter and nitrogen content in the influent,
in preparation for the detailed liquid and air measurement
campaign at the WWTP. For the initial sampling the
following parameters need to be monitored regarding the
activated sludge process:
• Influent flow rates (minimum of once per hour).
• Influent and effluent ammonia (up to eight times per
day or continuously).
• Influent and effluent nitrite and nitrate (up to eight
times per day or continuously).
• Influent and effluent COD (start with once per hour,
can be reduced depending on observed variability).
The strong relation between N2O emissions and
operational conditions makes the development of a
preliminary reconnaissance analysis crucial. The field
Additionally, diurnal performance and in-tank
profiles should be gathered at the time of the N2O gas
phase sampling. As far as feasible, all the liquid phase
184
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
from the online analysers present at a plant, to avoid
duplication of data-gathering efforts. In addition to the
data presented in Table 4.1, the following diurnal
performance and in-tank profiles should be gathered
(Table 4.2).
analyses must be according to approved methods and
protocols (e.g. APHA et al., 2012). Given the important
effect of nitrite on N2O emission, it is advised to give this
measurement special attention during the sampling
campaign. As far as possible, the sampling team should
work with the plant laboratory personnel to include data
Table 4.1 Overview of the data requirements for the preliminary WWTP assessment.
Sample location
Anoxic zone mixing
Influent flow
Parameter and sampling frequency (number of samples per week)
TSS
VSS
cBOD5,tot.1
cBOD5,sol.
CODtot.1
CODsol.
CODfilt,floc. Temp.
TKN1
TKNsol.
NH3-N
NO3-N
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
12
12
1
1
1
1
1
1
1
1
13
Sludge blanket TSS (use sludge judge 1/d and average 1/w)
Different measurements possible
• Approximate - set a settled sewage gate and allow for natural flow regime (if there is no information on the range of flows).
• Confirm the flow split by carrying out mass balance and MLSS concentration measurements.
• Alternatively, take a measurement of activated sludge at each pass.
• Use an optical density meter to get each pass TSS every 2-3 hours to get the moving average.
Mechanical or aerator-driven.
Diurnal flow pattern at appropriate time intervals (15 min for periods of rapid diurnal increase, 1 h for stable periods).
RAS flow
WAS flow
Dissolved oxygen
Aeration rate
Chemicals dosage
Average weekly RAS flow, indicate the location and type of flow measurement and variability of flow.
Average weekly WAS flow, indicate the location and type of flow measurement, times of WAS wasting (if it is not continuous).
1/d (then average 1/w), indicate the location of the DO measurement along the basin length and time of measurement.
Daily average, indicate the location of the airflow measurement and variability over the course of the day. SCADA output at short time intervals is preferred.
Daily, indicate the ferric chloride equivalent (and other chemicals) strength, dosing points and dose at each point
Settled sewage
Secondary effluent
Reactor MLSS
RAS MLSS
WAS MLSS
Clarifier
Flow split and flow rate
NO2-N
1
1
1
Homogenize subsample prior to the ‘total’ measurement. Discard the remaining sample – do not use for ‘filtrate’ or ‘soluble’ determinations.
Soluble COD can be used instead of filtered flocculated COD on the secondary effluent.
3
When RAS and WAS are from the same stream, a TSS measurement on one of these streams is sufficient.
RAS: return activated sludge; WAS: waste activated sludge.
2
Table 4.2 Complementary data requirements preferentially accompanying off-gas measurements.
Sample location
Settled sewage
Secondary
effluent
RAS MLSS
WAS MLSS
TSS
8
8
VSS
2
Parameter (number of samples per day)
cBOD5,tot.1 cBOD5,sol. CODtot.1 CODsol. CODfil.floc.
TKN1 TKNsol.
8
8
8
8
8
8
8
pH
8
Alk
8
NH3-N
8
NO3-N
8
NO2-N
8
82
8
8
8
8
8
8
8
8
82
8
8
8
83
Influent flow
Diurnal flow pattern at appropriate time intervals (15 min for periods of rapid diurnal increase, 1 h for stable periods).
RAS flow
WAS flow
Dissolved oxygen
Aeration rate
In-tank profiles
Average daily RAS flow, indicate the location and type of flow measurement and variability of flow.
Average daily WAS flow, indicate the location and type of flow measurement, times of WAS wasting if not continuous.
1/h, indicate the location of the DO measurement along the basin length and time of measurement.
Daily average, indicate the location of the airflow measurement and variability over the course of the day (SCADA outputs are preferred).
TSS
VSS
pH
DO
ORP
Temp.
COD filt.floc.
Alk.
NH3-N
NO3-N
NO2-N
8/d
2/d
8/d
8/d
8/d
8/d
8/d
8/d
8/d
8/d
8/d
OFF-GAS EMISSIONS TESTS
The collection of conventional wastewater samples
for analysis of parameters in Table 4.2 should preferably
be conducted by facility personnel who usually collect
operational and compliance samples. In advance of each
sampling event, the team conducting the monitoring
campaigns should consult with the laboratory personnel
to ensure that the samples for the conventional
parameters are collected during the GHG monitoring
event to meet the requirements of both the research
design and the WWTP’s laboratory operating
procedures.
The plant’s sample handling and custody
requirements can be utilized as far as possible for each
field sampling campaign. To confirm the adequacy of the
procedures, the plant’s procedures for field sample
handling and chain of custody should be reviewed with
the project team approximately two weeks prior to the
full-scale testing. At that time, if modifications are
185
deemed necessary by the project team, they can be
defined and documented in the site-specific sampling
protocol.
4.3.7. Testing and measurements protocol
Table 4.3 provides examples for the sample location, the
chemical parameter, sample container, preservative and
holding time for wastewater samples to be collected
during the measurements. This example is a starting
point and can be customized and expanded as required
for the specific plant where the sampling is conducted.
For the full-scale field testing, the plant laboratory will
follow their specific laboratory standard operating
procedures for each parameter. Standard operating
procedures from participating laboratories must be
reviewed for adequacy and consistency prior to the
measurement campaigns.
Table 4.3 Examples of sampling specifications to accompany off-gas monitoring.
Measurement classification
Sample location
Sample volume1 (mL) Sample preservation Max. holding time
Type
Frequency
Sample equipment
pH
C
NA
NA, in situ
Reactor
NA
NA
None, online
COD: Colorimetric
I, C
2/7 d
35 mL glass vial
Reactor, effluent
8
4 oC
1d
NH3-N: Potentiometric (ISE)
I, C
2/7 d
200 mL glass bottle
Effluent
80
4 oC 2
1d
NO2—N: Spectrophotometric
I, C
2/7 d
200 mL glass bottle
Effluent
40
4 oC 2
2d
NO3—N: Potentiometric (ISE)
I, C
2/7 d
200 mL glass bottle
Effluent
40
H2SO4, pH < 2
28 d
C
1/7 d
Gas sampling assembly
Reactor headspace NA
NA
NA
N 2O
NOx
C
1/7 d
Gas sampling assembly
Reactor headspace NA
NA
NA
DO
C
NA
NA, in situ
Reactor
NA
NA
None, online
1
The tabulated sample volume is twice that required for routine duplicate analysis and is apportioned into two sample containers. The additional volume is collected to determine quality
control measures such as accuracy (analysis of spiked samples), precision (duplicate analysis) and to account for potential sample loss while handling or analysing.
C: continuous measurement; I: intermittent measurement; Frequency of measurement applies only to continuous measurements
2
Storage at 4 °C. Note: the biomass is removed from the sample via centrifugation at 3,500 g for 10 min. Biomass removal arrests further biochemical oxidation of NH4+-N and NO2--N.
Parameter
4.4 EMISSION MEASUREMENTS
plant. The former is the most common approach and is
therefore presented in detail in this chapter.
The various methods for off-gas measurements used
in the research and practice are summarized in Table 4.4.
When a treatment plant is covered, the concentrations can
be easily measured in the ventilation air. Combining the
concentration measurements with airflow measurements
by e.g. use of a pitot tube (Klopfenstein, 1998) gives a
direct estimation of the emitted fluxes (Daelman et al.,
2015). Most treatment plants are however directly open
to the air and the collection of off-gas is a problem that
needs special attention for representative sampling. This
can be solved by either using gas hoods for gas collection
or measuring in the gas plume downwind of the treatment
When off-gas measurements are difficult (e.g. in
open tanks or when surface aerators prevent the
placement of flux chambers), emissions can be calculated
from the measurement of dissolved CH4 and N2O
concentrations. These can be integrated with mass
transfer coefficient measurements in mass balance
calculations to estimate the emission fluxes (Foley et al.,
2010). For dissolved N2O, microelectrodes can be used
to measure the concentration in the water phase on line
(Foley et al., 2010, Section 4.7.1), but sampling and gas
chromatographic analysis is also possible. For this
purpose, sampling the liquid phase and salting out the
186
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
soluble gasses to analyse them as a gas has been shown
to be an accurate and reliable method (Daelman et al.,
2012; Section 4.7.3).
Concentrations of dissolved gases such as N2O can be
continuously measured based on gas phase
measurements, according to a gas-stripping method
proposed by Mampaey et al. (2015) (Section 4.7.4)
Table 4.4 Comparison of different off-gas measurement techniques and approaches.
Disadvantages
Needs covered tanks. Difficult to use for
integral emission loads of a treatment plant, with spatial variations inside the plant.
Method
Liquid phase
Gas phase
Advantages
Direct measurements from
Sampling depending on
Continuous
Well-developed and easy to apply method. Gives
off-gas in covered tanks
accessibility to treatment tanks
temporal variations.
Flux chamber for open-surface Grab sample, N2O sensor,
process tanks
Continuous
Well-developed and widely applied protocols.
or using a stripping device with monitoring
Mechanisms and processes contributing to
gas measurement
emissions can be inferred based on spatially and
Multiple measurements needed to address
spatial and temporal variability and to
obtain full emission loads of a plant.
temporally based measurements (also zonespecific measurements).
Measuring the emission
plume downwind of the
treatment plant
Not sampled
Continuous
A skilled operator is required.
Dependence on favourable wind
Downwind plume changes can be instantaneously conditions combined with road access.
Monitoring only possible with favourable
detected and the measurements adjusted
wind.
accordingly.
Flexibility to carry the equipment around either by Very limited spatial resolution.
Straightforward data analysis and calculation
when gases are fully mixed.
car or small trolley.
Capability to point out emissions from hotspots.
Whole plant emission quantification.
A third alternative for measuring the emission of a
total wastewater treatment plant is the detection of the
CH4 and N2O concentrations in the downwind plume of
the plant. This needs very specialized equipment such as
photoacoustic or cavity down-ring spectroscopy
(Yoshida et al., 2014, Rapson et al., 2014) and is
therefore only usable by specialists. The advantage is that
monitoring can be done outside plant boundaries, and the
full plant with all its process units is monitored. Hotspots
at the plant can be identified (e.g. influent works for
methane emission or ventilation air stack); however,
accurate spatial resolution for e.g. the aeration tanks is
not possible. Given the costs of the equipment, long-term
monitoring is economically often not possible. The
method is established for landfill monitoring, where
emissions are estimated by sampling a few days per year
with mobile equipment.
4.5 N2O MEASUREMENT IN OPEN TANKS
The overall procedure for measuring off-gas fluxes from
the headspace of open surface activated sludge tanks can
involve a flux-chamber approach, which has been used
by most published work to date. This works for most
installations, except for surface-aerated systems. When
the surface-aerated space is covered, the gas composition
in the ventilation air (and its flow rate) can be measured.
When the aerators are open in the air, one has to rely on
the use of liquid phase balances for N2O or CH4 (Foley et
al., 2010; Section 4.7).The principal advantage of the
flux chamber approach is that it can yield some
understanding of the spatial variability in emissions
across different zones or the reactor of an activated
sludge p rocess. On the other hand, it is operationally
intensive to conduct due to the need for multiple
sampling locations to get representative measurements.
In the US, the off-gas measurement method is a variant
of the EPA/600/8-86/008 and the South Coast Air
Quality Management District (SCAQMD) tracer
methods. This variant was developed to measure sources
that have a relatively high surface flux rate when
compared to diffusion (for instance, spilled oil
containment).
Commercially available replicas of the US EPA
surface emission isolation flux chamber (SEIFC, figures
4.1-4.3) are used to measure gaseous nitrogen fluxes
from activated sludge reactors. The SEIFC consists of a
floating enclosed space from which exhaust gas can be
OFF-GAS EMISSIONS TESTS
collected as a continuous or grab sample. Because the
surface area under the SEIFC can be measured, the
specific flux of the gaseous compound of interest can be
indirectly determined. The SEIFC ‘floats’ on the
activated sludge tank surface, and several replicate
measurements can be taken at different locations in a
single tank as well as from different tanks (nitrification,
denitrification) along a treatment train.
The SEIFC is equipped with mixing (a physical mixer
or via sweep gas circulation) to ensure adequate gas
mixing and, ideally, an online temperature probe. The
SEIFC is currently one of the few devices accepted by
the US EPA for measuring gaseous fluxes (Tata et al.,
2003). Continuous gas-phase analyses are conducted via
infrared (N2O/CH4). Although not discussed in this
chapter, the gas could also be measured for oxygen and
carbon dioxide for measuring bio-conversions in the
wastewater treatment process.
In general, sampling in open surface tanks needs to
be conducted at multiple locations of the activated sludge
train in each wastewater treatment facility. These
locations include aerobic, anoxic, and anaerobic zones,
depending upon the configuration of the given facility. In
non-aerated zones a sweep gas is used to produce an
effective gas flow through the flux chamber. During the
course of the gas phase sampling, liquid phase samples
for dissolved compounds such as ammonium, nitrate and
nitrate also need to be collected adjacent to the hood
location. The samples should be filtered immediately
upon collection in the field and can be analysed utilizing
readily available field methods (i.e. Hach kits) and
standard laboratory analytical methods (APHA et al.,
2012). This allows the measured gas data to be linked
with the actual concentrations of relevant compounds at
the point of measurement. Comparing the data with
influent/effluent measurements or measurements at other
spots in the tank might give biased correlation between
the off-gas and liquid phase samples.
In open surface tanks, the specific locations selected
are typically close to the influent or effluent end of each
demarcated anoxic or aerobic zone in the WWTP.
Continuous measurement at each of these specific
locations is typically conducted over a minimum 24-h
period or longer.
The treatment trains of selected wastewater treatment
plants that accomplish nitrification and denitrification are
characterized based on their liquid-phase and gas-phase
nitrogen concentrations and speciation. Testing is
187
conducted at each location during a sampling campaign
during which gas-phase monitoring is conducted in realtime continuous mode and liquid-phase sampling is
conducted via discrete grab sampling. Trends and
variations in gaseous emissions and speciation are also
ascertained. This sampling effort is intended to assist in
the development of process-operating criteria that
minimize both gaseous and liquid-phase nitrogen
emissions from wastewater treatment facilities. The
analysis of nitrogen GHG compounds and precursors in
both the air and liquid phases is complemented by
analysis of conventional wastewater parameters.
Monitoring of the liquid phase and the gas phase can
be conducted in seasonally distinct regimes, for instance,
once in warm temperature conditions (i.e. summer, early
autumn), and cold temperature conditions (winter/early
spring) to account for annual variability.
The overall procedure for measuring N2O and CH4
fluxes from the headspace of activated sludge tanks
involves a variant of the EPA/600/8-86/008 and the
SCAQMD tracer methods. Gas-phase analyses can be
conducted via infrared (N2O/CH4) analysers.
In the absence of an approved method for N2O in air
or water, method modification can be necessary to
measure N2O emissions. In a recent study (Ahn et al.,
2010a), in order to evaluate the performance of the
measurement of N2O fluxes using the procedure
developed by the researchers, three side-by-side
monitoring events were conducted along with the
research procedure during the first sampling event at a
step feed BNR facility. In addition to the standard
research protocol, two additional side-by-side
monitoring events were conducted as follows (Table 4.5).
Plant wastewater research engineers measured fluxes
using the EPA isolation flux chamber and SCAQMD
tracer method and with a photo acoustic analyser to
directly determine N2O. Plant consultants used the
textbook EPA isolation flux chamber and SCAQMD
tracer dilution method to measure the flux and the
following analytical methods to measure ozone
precursors and GHGs. These side-by-side tests were not
designed to validate the modified analytical approach to
establish an approved methodology. However, they
provided an independent verification that the approach
accurately measured nitrogen GHG emissions to meet the
objectives of that study, for zones where concurrent sideby-side measurement was conducted. This study was part
of the original Water Environment Research Foundation
188
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
(WERF) project, during which this protocol was
developed.
Table 4.5 Additional comparative analytical methodologies for measuring
off-gas emissions from wastewater treatment plants.
Method/Species
ASTM method 1946permanent gas
analysis
NIOSH 6600
Technique
GC/TCD
FTIR
Application
Relevant fixed gases: CH4, CO,
CO2, and helium (He) as a
separate analysis.
N2O
Based on this side-by-side comparison, it was further
recommended that subsequent studies should consider
the helium (He) tracer method (based on ASTM D1946)
to measure the gas flow rate from the flux chamber, the
results of which are reported in (Ahn et al., 2010b).
4.5.1 Protocol for measuring the surface flux
of N2O
The following protocol is intended to provide researchers
and field sampling teams with a detailed description of
the data collection methodology and analysis
requirements to enable the calculation of gaseous
nitrogen fluxes from different zones of activated sludge
trains in a WWTP.
4.5.1.1 Equipment, materials and supplies
The following equipment is needed to perform the
protocol; suppliers and manufacturers may vary:
1. A surface emission isolation flux chamber
(commercially available from vendors or custombuilt based on specifications from the US EPA
(Kienbusch, 1986)).
2. Teledyne API N2O Monitor Model 320E (Teledyne
API, San Diego, CA).
3. Zero gas (containing zero ppm N2O and CH4), and
N2O and CH4 gas standards (Tech Air, White Plains,
NY).
4. Dwyer series 475 Mark III digital manometers to
measure the flux chamber pressure from 0 to 1” (2.54
cm water column) (high sensitivity) and 0 to 100”
(low sensitivity) of a water column (Dwyer
Instruments Inc., Michigan City, IN).
5. A rotameter to measure the influent sweep gas flow
rate, 0-30 L min-1 (Fisher Scientific, Fairlawn, NJ).
6. An adjustable air pump, 0-10 L min-1 (Fisher
Scientific, Fairlawn, NJ) to provide sweep gas flow
into the flux chamber.
7. A vacuum pump, 0-30 L min-1 (Fisher Scientific,
Fairlawn, NJ) for active pumping of gas from the flux
chamber, if needed.
8. 0.2 μm cartridge filters, set of 10 (Millipore, Ann
Arbor, MI) to prevent fine particulates from entering
the gas analysers.
9. A silica gel column for capturing moisture (Fisher
Scientific, Fairlawn, NJ).
10. A glass water trap consisting of a 100 mL glass bottle
placed in ice within a Styrofoam® box.
11. Teflon® tubing (approximately 0.5”) and fittings.
12. A 100-300’ extension cord and power strip.
13. A laptop personal computer (with at least 512 MB
RAM) with data acquisition programs for N2O and
CH4 analysers pre-installed.
14. A set of miscellaneous hand tools including
adjustable wrenches, different size screwdrivers and
adjustable pliers.
4.5.1.2 Experimental procedure
The overall procedure for measuring N2 O and CH4 fluxes
from the head-space of activated sludge tanks involves a
variant of the EPA/600/8-86/008 and the South Coast Air
Quality Management District (SCAQMD) tracer
methods, which allow sampling of gaseous emissions
from high surface flux rate operations.
This variant has been developed to measure those
sources that have a relatively high surface flux rate when
compared to diffusion; this facilitates increased sampling
at composting and wastewater treatment plants.
Commercially available replicas of the US EPA
surface emission isolation flux chamber (SEIFC) are
used to measure gaseous N fluxes from activated sludge
reactors. The US EPA SEIFC essentially consists of a
floating enclosed space through which carrier gas
(typically nitrogen or argon) is fed at a fixed flow rate
and exhaust gas is collected in a real-time or discrete
fashion. Since the surface area under the SEIFC can be
calculated or measured, the specific flux of the gaseous
compound of interest can thus be determined. Since the
SEIFC ‘floats’ on the activated sludge tank surface,
several replicate measurements can be taken at different
locations in a single tank as well as from different tanks
(nitrification, denitrification) along a treatment train. The
SEIFC is also equipped for mixing (with a physical mixer
or via sweep gas circulation) to ensure adequate gas and
OFF-GAS EMISSIONS TESTS
189
in some cases, an online temperature probe. The SEIFC
is currently one of the few devices accepted by the US
EPA for measuring gaseous fluxes (Tata et al., 2003).
Gas-phase analyses are conducted via infrared (N2O,
CH4) methods.
Gas
Out,Qflux
To pressure sensor
Sweep air
In,Qsweep
To analyzers
Gas from AS
tanks,QA/S
Figure 4.1 Schematic of the US EPA surface emission isolation flux
chamber (modified from Tata et al., 2003).
Pressure build-up can be minimized by equipping the
flux chamber with multiple vents or a variable size vent
and continuously monitoring the pressure drop across the
hood using a sensitive pressure gauge. In all the field
locations, the gas flow rate should be measured using the
tracer gas technique and pressure across the flux chamber
should be passively monitored if necessary.
Alternatively, the aeration rate from plant records
(available as an order of magnitude verification) can be
used. The modified setup of the flux chamber used in this
study is depicted in figures 4.1-4.3.
During the course of the gas phase sampling,
liquid phase samples are collected adjacent to the
hood location. The samples should be filtered
immediately upon collection in the field and analysed by
the host plant personnel for ammonia, nitrite and nitrate
concentration, utilizing readily available field methods
(i.e. a Hach Kit). As the primary purpose of these
measurements is to ensure the presence of the targeted
nitrogen species, without consideration for the accuracy
in the concentration measurements, the simplest
available field method will be used for these preliminary
measurements.
To pressure sensor
Sweep air in
In general, sampling is conducted at multiple
locations of the activated sludge train in each wastewater
treatment facility. These locations are the aerobic, anoxic
and anaerobic zones, depending upon the configuration
of the given facility. Additionally, within each zone,
multiple points (approximately three, but not less than
two) must be sampled to address any variability in gas
fluxes that may result from variations in mixing or flow
patterns therein.
Sweep air out
Sample ports
4.5.1.3 Sampling methods for nitrogen GHG
emissions
Figure 4.2 Modified schematic of the flux chamber.
The gas-phase sampling method in aerobic zones
Online N2O
analyzer
Temperature
monitor
Moisture
trap
Rotameter
Gas out
Online NOX
analyzer
0.2 µm
cartridge filters
Moisture
trap
Pressure
gauge
Sweep air in
Figure 4.3 Schematic of flux-chamber set-up for gas flux measurements.
1. Seal all but one vent in the flux chamber and connect
a high-sensitivity pressure gauge to the one open
vent.
2. Lower the flux chamber into the aerobic zone (the
bottom of the rim should be below the surface of the
water by 2.5-5.0 cm minimum).
3. Wait for the N2O analyser to equilibrate, based on the
stability indicator (<0.03)
4. Pull the flux chamber up. Open two vents and
connect the gas analyser. The other vents should be
left open to the atmosphere.
190
5. Record the temperature of the gas in the flux chamber
using a digital temperature gauge (Fisher Scientific
number 15-077-8 or a suitable alternative).
6. Care must be taken that the flow going to the two
analysers does not exceed the gas flow rate from the
flux chamber. Otherwise, atmospheric air will be
drawn in through the vents in the flux chamber.
Determination of the gas flow rate from the flux chamber in
aerobic zones
1. Disconnect the gas analysers and connect one outlet
vent to the inlet line of a field gas chromatograph
equipped with a thermal conductivity detector. Close
the other vent.
2. Introduce tracer gas (10 % helium, 90 % zero air)
through an inlet vent into the flux chamber at a known
flow rate (for instance 1 L min-1).
3. Measure the concentration of helium gas exiting the
flux chamber (see the protocol in Section 4.6).
4. Based on the measured helium concentrations,
calculate the flow rate of the aeration tank headspace
gas entering the flux chamber (Eq. 4.1, Section 4.6).
The gas-phase sampling method in anoxic zones
1. Seal all but one vent in the flux chamber and connect
a high-sensitivity pressure gauge to the one open
vent.
2. Lower the flux chamber into the anoxic zone with a
1-2 inch minimum submergence, into the liquid
surface.
3. Wait for the N2O analyser to equilibrate, based on the
stability indicator (< 0.03)
4. Pull the flux chamber up. Open two vents and
connect the gas analyser and the sweep gas pump
(Note: sweep gas is only used during anoxic zone
sampling). The other vents should be left open to the
atmosphere.
5. Record the temperature of the gas in the flux chamber
using a digital temperature gauge (Fisher Scientific
number 15-077-8 or a suitable alternative).
6. Care must be taken never to allow the flow going to
the two analysers to exceed the sweep gas rate, or
dilution air will be drawn in through an opening in
the chamber.
Determination of the gas flow rate from the flux chamber in
the anoxic zone
1. Disconnect the gas analysers and connect one outlet
vent to the inlet line of a field gas chromatograph
equipped with a thermal conductivity detector. Close
the other vent.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
2. Introduce sweep gas to the chamber at a flow rate of
4 L min-1 and wait 6 min for a steady state.
3. Introduce tracer gas (10% He, 90% zero air) through
an inlet vent into the flux chamber at a known flow
rate (for instance 1 L min-1).
4. Measure the concentration of He gas exiting the flux
chamber.
5. Based on the measured He concentrations, calculate
the flow rate of aeration tank headspace gas entering
the flux chamber (Eq. 4.1, Section 4.6).
Table 4.6 summarizes the data recording requirement
checklist that needs to be followed for the flux-chamber
setup and operation. Additional parameters can be added
by sampling teams based on a case-specific basis.
Table 4.6 Checklist for flux chamber set-up and operation in field.
Measurement
Sampling
location 1
Sampling
location 2
Sampling
location 3
Pressure in flux
chamber
Gas flow rate from
flux chamber
Gas temperature in
flux chamber
Wastewater
temperature
Air-pump
flow rates
Continuous and real-time gas measurement
1. Turn on the power by pressing the on/off switch on
the front panel. The display should turn on and the
green (sample) status LED should blink, indicating
the instrument has entered the HOLD-OFF mode.
Sample mode can be entered immediately by pressing
the EXIT button on the front panel. The red ‘fault’
light will also be on until the flows, temperatures and
voltages are within operating limits. Clear the fault
messages. After the warming up, review the TEST
function values in the front panel display by pushing
the button TEST on the left of the display.
2. Activate the DAS data acquisition software and set
the sampling frequency for 1 sample per minute.
3. Start the data acquisition.
4. Connect the inlet tubing of the analyser to the outlet
tubing from the SEIFC securely using a standard 1/4”
compression fitting connector.
5. Acquire data for about 20 min in anoxic zones and
about 10 min in aerobic zones after stable readings
OFF-GAS EMISSIONS TESTS
6.
7.
8.
9.
are obtained, as indicated by the stability indicator on
the N2O analyser.
Stop the DAS software and immediately save the
acquired data.
Repeat steps 2-5 for each sampling point and
sampling locations (individual tanks).
Note that the measurement range is 0-1,000 ppm.
Before each sampling event, calibrate the instrument
by using ‘zero gas’ and N2O standard gas as per the
manufacturer’s instructions.
Principles of real-time N2O and CH4 measurements
Continuous N2O and CH4 measurements are performed
via infrared (IR) gas-filter correlation, which is based on
the absorption of IR radiation by N2O and CH4 molecules
at appropriate wavelengths. As part of the measurement
process, a broad wavelength IR beam is generated inside
the instrument and passed through a rotating gas filter
wheel, which causes the beam to alternatively pass
through a gas cell filled with nitrogen (a measure cell),
and a cell filled with N2O/N2 mixture (a reference cell) at
a frequency of 30 cycles per second. N2O concentrations
are inferred based on the amount of IR absorption.
Ultimately, the ‘stripped’ beam strikes the detector,
which is a thermo-electrically cooled solid-state photoconductor. This detector, along with its pre-amplifier
converts the light signal into a modulated voltage signal.
4.5.1.4 Direct measurement of the liquid-phase N2O
content
In addition to measuring gaseous phase N2O
concentrations in the headspace of aerobic and anoxic
zones, liquid phase N2O concentrations could also be
measured to discriminate between N2O generation in the
liquid phase and N2O emission in the gas phase. Liquidphase N2O concentrations can be measured using a
polarographic Clark-type electrode (Unisense, Aarhus,
Denmark). For additional details of the liquid phase
measurements summarized in this section, please refer to
the protocols in Section 4.7.
1. Withdraw a sample of about 20 mL from the test
reactors in 50 mL conical centrifuge tubes or
alternatively similar containers (plastic or glass
beakers are acceptable).
2. Take out the microsensor from the calibration
chamber (containing deionized water), rinse out with
deionized water, and mop dry with a tissue.
3. Immerse the microsensor into the samples. Proceed
as rapidly as possible after obtaining the sample.
191
4. Record the numbers from the display on the
picoammeter. The measurement numbers should
become stable within one minute.
5. Pull out the microsensor, rinse out and place it back
into the calibration chamber.
6. Repeat steps 1 to 5 for each sampling point and
location.
The overall sampling effort is rather involved and
thus needs to be closely coordinated. The real-time data
from the analysers or probes should be automatically
downloaded onto a field computer or recorded in
laboratory notebooks. All the electronic data should be
backed up frequently and where feasible electronic data
should be stored on a temporary disk drive (in addition to
the PC hard drive) during the field testing events.
4.6 MEASUREMENT OF OFF-GAS FLOW IN
OPEN TANKS
This section describes a protocol for the measurement of
off-gas flow rates using a helium tracer gas method (after
ASTM method D1946). While measuring and reporting
gaseous emission fluxes or while performing mass
balances, the measurement of both headspace gas
concentrations and the overall off-gas flow rates is
needed. While gaseous N2O and CH4 concentrations
from the activated sludge surface can be measured
online, it is often not practical to conduct online or even
discrete
round-the-clock
advective
gas
flow
measurements. While blower operational data can be
used as an approximation for determining advective flow,
however usually only run time or electricity use are
measured and conversion to actual air flow has
uncertainties. Moreover, this approach suffers some
serious uncertainties in the link between overall bulk
blower air flow and actual distribution among different
aerated zones. Direct measurement of the gas flow in
aeration air pipes is an option, or for closed tanks often
the off-gas flow rate can be measured. For open tanks a
helium tracer method can be used, but not in a continuous
mode owing to cost and plant access issues. At this point,
the only way to overcome this limitation for open tank
systems and reduce the inherent variability from
extrapolating flow data from a few discrete
measurements is to measure the advective flow rate using
the He tracer method more frequently and employ a
correlation analysis to translate the discrete flow
192
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
measurements through the flux chamber to continuous
tank-scale measurements.
4.6.1 Protocol for aerated or aerobic zone
Advective flow of gas through the flux-chamber
(Qemission) in aerated zones is measured using a
modification of the ASTM method D1946. Briefly, a
tracer gas consisting of 100,000 ppmv (Chelium-tracer)
helium is introduced into the flux-chamber at a known
flow rate, Qtracer (Eq. 4.3). Helium concentrations in the
off-gas from the flux chamber (Chelium-FC) can be
measured using a field gas chromatograph equipped with
a thermal conductivity detector (GC-TCD). Qemission can
then be computed using Eq. 4.1.
Protocol steps
1. Activate the field gas chromatograph approximately
prior to the actual helium measurements to allow for
the thermal conductivity detector (TCD) and GC
column to attain the desired temperatures.
2. After measuring the gas-phase concentrations,
disconnect the gas analysers and connect one outlet
vent to the inlet line of the field GC. Close the other
vent.
3. Introduce tracer gas (10 % helium, 90 % zero air)
through an inlet vent into the flux chamber at a known
flow rate (for instance 1 L min-1).
4. Measure the concentration of helium gas exiting the
flux chamber (as per the ASTM method D1946).
5. Based on the measured helium concentrations,
calculate via linear algebra the flow rate of the
aeration tank headspace gas entering the flux
chamber (Eq. 4.1).
Q tracer ⋅ C helium − tracer = (Q tracer + Q emission ) ⋅ C helium − GC
Q emission =
Q tracer ⋅ (C helium − tracer − C helium − GC )
C helium − GC
Eq. 4.1
6. For each sampling location, repeat steps 2-5 at least
three times.
4.6.2 Protocol for non-aerated zones
The only modification to the protocol to measure the
emission flow rate from non-aerated zones is the
introduction of sweep gas (air) or carrier gas through the
flux chamber at a known flow rate (Qsweep), in addition to
the helium tracer gas. The corresponding Qemission is
computed using Eq. 4.2. The addition of sweep gas is
needed to promote mixing of the SEIFC contents, owing
to the low advective gas flow from the anoxic-zone
headspace. Sweep-air N2O and CH4 concentrations
always need to be measured and checked to ensure that
they remain below the detection limits of the N2O or CH4
analysers.
Protocol steps
1. The only modification to the protocol for adaptation
to measuring the emission flow rate from the anoxic
zone is the introduction of sweep gas.
2. Introduce sweep gas to the chamber at a flow rate of
4 L min-1 and wait 6 min for a steady state.
3. Follow steps 2-6 as described above for
determination of the emission flow rate from aerobic
zones.
4. Calculate the emission flow rate from the anoxic zone
using Eq. 4.2.
Q tracer ⋅ C helium − tracer = (Q tracer + Qsweep + Q emission ) ⋅ C helium − GC
Qemission =
Q tracer ⋅ (C helium − tracer − C helium − GC )
− Qsweep
C helium − GC
Eq. 4.2
Each sampling campaign consists of discrete and
continuous N2O measurements. During the discrete N2O
measurements, Qemission should be determined at each
location in the treatment plant where N2O is measured.
During continuous N2O measurements, Qemission should be
determined several times a day in correspondence with
liquid-phase measurements.
In the case of surface-aerated tanks measuring actual
gas flow rates, the actual fluxes of greenhouse gasses is
usually an even more difficult task. In these cases
emissions can be best estimated from liquid-phase
measurements and balances over the liquid phase.
4.7 AQUEOUS N2O and CH4
CONCENTRATION DETERMINATION
The aqueous concentrations of nitrous oxide can be
directly measured by oxygen electrodes that are polarized
differently than for oxygen measurement. For strict
anaerobic or anoxic conditions (i.e. no oxygen present)
these adapted electrodes can be directly used. For
measuring N2O concentrations in an aerated aqueous
phase, a miniaturized Clark-type sensor with an internal
reference and a guard cathode is available, and can be
successfully applied (Foley et al., 2010). The sensor is
equipped with an oxygen front guard to prevent oxygen
from interfering with the N2O measurements. This
OFF-GAS EMISSIONS TESTS
method is described in Section 4.7.1. For methane no
dissolved gas-phase measurement is available.
The concentrations of N2O and CH4 in a liquid
sample can be determined by driving the dissolved gas
into the gas phase. In a grab sample this can be done by
salting out the gasses into the headspace of the sample
tube, and analysing the gas phase on a GC (Section 4.7.3).
With a gas-stripping device it is possible to continuously
monitor the dissolved gas concentrations (Section 4.7.4).
4.7.1 Measurement protocol for dissolved
N2O measurement using polarographic
electrodes
4.7.1.1 Equipment
1. A nitrous oxide microsensor N2O25 (Unisense,
Aarhus, Denmark).
2. A two-channel picoammeter PA2000 (Unisense,
Aarhus, Denmark).
3. A calibration chamber CAL300 (Unisense, Aarhus,
Denmark).
4. Zero air and N2O gas standard (Tech Air, White
Plains, NY).
5. Teflon® tubing, silicone tubing and fittings.
6. A squeezer with deionized water.
7. Kimwipes.
8. BD Falcon 50 mL conical tubes.
4.7.1.2 Experimental procedure
The Unisense nitrous oxide microsensor is a miniaturized
Clark-type sensor with an internal reference and a guard
cathode. In addition, the sensor is equipped with an
oxygen front guard, which prevents oxygen from
interfering with the nitrous oxide measurements. The
sensor is connected to a high-sensitivity picoammeter
and the cathode is polarized against the internal
reference. Driven by the external partial pressure, nitrous
oxide from the environment will penetrate through the
sensor tip membranes and be reduced at the metal
cathode surface. The picoammeter converts the resulting
reduction current into a signal. The internal guard
cathode is also polarized and scavenges oxygen in the
electrolyte, thus minimizing zero-current and prepolarization time.
193
The measurement steps are as follows:
a. Turn on the power switch located on the front panel
of the picoammeter.
b. Check that the ‘Gain’ screw for Channel 1 is turned
fully counter-clockwise.
c. Turn the display switch, located on the centre of the
panel, to ‘Signal 1’ and check that the display reads
zero. If not, adjust the offset, as per the
manufacturer’s instructions.
d. Turn the display switch to ‘Pol. 1’. Check if the
polarization voltage shows -0.8 V. If not, adjust the
volt and polarity switch.
e. Connect the leads of the ‘pre-polarized’ microsensor
to the meter in the following order: (1) Signal wire
(black) to ‘Input’ of Channel 1 on the front panel. (2)
Guard wire (yellow) to ‘Guard’ of Channel 1.
f. Rinse out the sensor with deionized water and absorb
the moisture with tissue paper.
g. Place the sensor into the calibration chamber, which
contains deionized water.
h. Select the ‘Normal’ setting for the ‘Mode’ switch on
the front panel, unless you need the extremely fast
response.
i. Select the appropriate measuring range using the
‘Range’ switch on the panel. Usually 200 pA is
selected, but if not suitable, select an alternative
available range.
j. Withdraw about 20 mL sample from test reactors in
50 mL conical centrifuge tubes or alternatively
similar containers (plastic or glass beakers are
acceptable).
k. Take out the microsensor from the calibration
chamber (containing deionized water), rinse out with
deionized water, and mop dry with a tissue.
l. Immerse the microsensor into the samples. For (j) and
(k), proceed as rapidly as possible after acquiring the
sample.
m. Record the numbers from the display on the
picoammeter. The measurement numbers should be
stable within one minute.
n. Pull out the microsensor, rinse out and place it back
in the calibration chamber.
o. Repeat steps (j) to (n) for each sampling point and
location.
p. When the measurements are complete, disconnect the
sensor leads in the reverse order to which they were
connected.
The measurement range is adjustable, 0-0.616
ppmv-N2O (with 500 ppm N2O gas standard).
194
If the sensor is new or has not been operated for
several days, then it must be polarized for at least 2 h and
up to 12 h before it can be calibrated and/or used, as
follows:
a. Secure the nitrous oxide sensor with its tip, immersed
in nitrous oxide-free water.
b. Turn the display switch to ‘Pol. 1’ and adjust the
polarization to -1.30 V.
c. Turn the display switch to ‘Signal 1’ and adjust the
‘Gain’ screw completely counter-clockwise. Adjust
the display to zero on the ‘Offset’ dial, if needed.
d. Connect the signal wire (black) of the microsensor to
the ‘Input’ terminal.
e. After 5 min, adjust the polarization to -0.8 V and then
connect the guard wire (yellow) to the ‘Guard’
terminal.
f. Pre-polarize for up to 12 h if possible to get the
maximum stability.
After the sensor has been polarized, it must be
calibrated with zero air and N2O gas standards. Typically,
we have used 500 ppm N2O gas standards for calibration.
Note that N2O gas standards are specialised items and can
be purchased from vendors such as TechAir.
To be consistent in terms of units for liquid and gas
phase N2O, the results of this study are expressed in terms
of N2O. Alternatively, liquid and gas phase N2O
concentrations can also be expressed as ‘N’ to estimate
the fraction of influent nitrogen discharged as N2O.
4.7.2 Measurement protocol for dissolved
gasses using gas chromatography
Both nitrous oxide and methane can be readily analysed
by standard gas chromatographic analysis (Weiss, 1981).
For liquid samples this can be performed by taking a
sample and transferring it directly to a closed bottle or
tube with a calibrated volume. In the bottle an
equilibrium will establish for the volatile compounds
between gas and liquid phase. After equilibration the gas
phase concentration can be measured. With the Henry
coefficient (corrected for temperature and ionic strength)
the liquid concentration for methane or nitrous oxide can
be calculated. Using these concentrations and the known
gas and liquid volumes in the sample bottle then gives the
total mass of compound in the bottle and the original
concentration in the sample. This method is prone to
many measurements and uncertainties for e.g. the true
Henry coefficient for the actual fluid composition.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
It is therefore more reliable to drive all the soluble
gasses into the gas phase by e.g. a high salt concentration
(the salting-out method). This has been described in
detail by Daelman et al. (2012) based on the salting-out
method described by Gal’chenko et al., (2004). The
measurement protocol is described below.
4.7.3 Measurement protocol for dissolved
gas measurement by the salting-out
method
Before the start of each sampling round, serum bottles of
120 mL are filled with 20 g NaCl. At the different
sampling locations at the WWTP, samples are collected
with a sampling beaker. From this beaker, 50 mL of the
sample is added carefully to a serum bottle filled with
salt, using a syringe with a catheter tip and a 10 cm
silicone tube. While emptying the syringe into the bottle,
the silicon tube is held under the rising liquid surface in
order to keep the liquid-gas interface as small as possible
to avoid stripping. Immediately after adding the syringe
content to the bottle, the bottle is sealed with a rubber
stopper and an aluminium cap. The sealed bottle is
shaken vigorously in order to speed up the dissolving of
the salt. At 20 °C the solubility of NaCl in water is about
360 g L-1, so these samples, containing 400 g NaCl L-1,
are oversaturated. As a consequence of the high salt
concentration, the microbial activity in the sludge
samples is halted, and the dissolved gases are salted out.
Dissolved methane, but also the other dissolved gases
such as carbon dioxide escape from the liquid phase to
the headspace of the serum bottles. This results in a
pressure build-up in the headspace. Before sampling the
headspace of the samples for analysis with GCFID, the
pressure in the headspace needs to be equilibrated with
the atmosphere by allowing the gas in the headspace to
expand in a submerged graduated syringe. The increase
in gas volume is used to calculate the pressure build-up
in the headspace. After the gas pressure in the headspace
is brought to atmospheric pressure, the headspace is
sampled with a gas syringe and analysed in the
conventional procedure by gas chromatography. The
amount of methane in the headspace before expansion is
calculated from the concentration, the measured volume
of the headspace of the sealed bottle and the headspace
pressure after the pressure build-up that is calculated
from the volume expansion of the headspace. Before the
sample was saturated with salt, this amount of methane
in the gas phase must have been completely dissolved in
the liquid sample. By dividing this amount by the sample
OFF-GAS EMISSIONS TESTS
volume (50 mL), the original methane concentration in
the liquid can be established.
195
h. Register the new headspace volume in the syringe V1.
i. V1 – V0 = Vs with Vs the volume expansion due to the
pressure build-up in the serum bottle.
4.7.3.1 Equipment
The equipment needed in this protocol is as follows:
• A sampling beaker.
• A graduated syringe with a catheter tip and tube of
ca. 15 cm.
• A serum bottle of ca. 120 mL.
• A rubber stopper.
• An aluminum seal.
• A crimper for aluminium seals.
• 20 g NaCl.
• A beaker.
• A graduated syringe (or burette) with a catheter tip
and without a plunger.
• A tube of ca. 30 cm.
• A hypodermic needle.
4.7.3.2 Sampling procedure
•
•
•
•
•
•
Before sampling starts, add 20 g NaCl to the serum
bottle.
Take a sample from the liquid surface of the reactor
or from the valve of a pipe using a sampling beaker.
Suck 50 mL (Vsample) of the sample from the sampling
beaker into the syringe with the catheter tip and tube.
Release the content of the syringe into the serum
bottle, while keeping the end of the tube under the
liquid surface.
Seal the serum bottle with the rubber stopper and the
aluminium seal.
Shake the bottle vigorously.
4.7.3.3 Measurement procedure
Measurement of the volume expansion due to pressure buildup in the bottle
a. Place the graduated syringe (or burette) in a beaker
with water.
b. Connect the graduated syringe (or burette) with the
catheter tip to the needle with a tube.
c. Make sure that the open end of the syringe is
submerged in the water in the beaker.
d. Register the headspace volume in the syringe V0.
e. Pierce the rubber stopper of the serum bottle with the
needle connected to the syringe (Figure 4.4).
f. The headspace volume of the syringe will expand.
g. Bring the water level in the syringe to the water level
in the beaker to cancel out the pressure of the water
column.
Figure 4.4 Measurement of the volume expansion due to the pressure
build-up that is caused by the salting-out of the dissolved gases (photo:
van Dongen, 2015).
Measurement of the methane or the nitrous oxide
concentration in the headspace
Draw a sample from the headspace of the serum bottle
with a gas syringe and measure it with a gas
chromatograph equipped with a flame ionization
detector, according to the appropriate method for
measuring methane.
Measurement of the headspace of the serum bottle
a. Mark the level of the NaCl saturated sample in the
serum bottle.
b. Mark the lower side of the rubber stopper.
c. Empty and rinse the bottle.
d. Fill the bottle with clean water up until the mark of
the liquid level.
e. Register the weight of the bottle W0.
f. Add clean water to the bottle up until the mark of the
stopper.
196
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
g. Register the weight of the bottle W1.
h. (W1 – W0) × ρ = VHS with ρ the density of the water
and VHS the volume of the headspace of the serum
bottle.
Concentration
C=
Volume
4.7.4 Measurement protocol for dissolved
gas measurement by the stripping method
Eq. 4.4
Where, V is the expanded volume of the headspace
(m³), Vs is the volume expansion due to the pressure
build-up (m³), and VHS is the headspace of the serum
bottle before expansion (m³).
Amount of methane
n=
4.7.4.1 Operational principle
This method was first published in Mampaey et al.
(2015). By using a gas-stripping device, it is possible to
monitor dissolved gases, e.g. N2O, on a continuous basis.
The unknown liquid concentration can be calculated
from the measured gas-phase concentration in the off-gas
of a stripping device. The proposed method relies on a
gas-stripping device, consisting of a stripping flask and a
scum trap flask, as displayed in Figure 4.5.
Eq. 4.5
P·V
R·T
Where, n is the amount of methane in the expanded
headspace of the serum bottle (mol), P is the atmospheric
pressure (Pa), V is the expanded volume of the headspace
(m³), R is the ideal gas constant: 8.314 m³ Pa mol-1 K-1,
and T is the temperature (K).
Gas outflow
reactor
N2 in
R (t)
CG,2
R (t)
CG,1
CRL (t)
Gas outflow
QG,
CG,2 (t)
VG,2
Subsystem 2:
Headspace
CG,2 (t)
QG,1(t)
QLR
R
VG,1
VRL
CL(t)
L
Liquid outflow
reactor
Reactor
CGin,R= 0
Liquid inflow
QL C R (t)
R
VG,2
R (t)
CG,2
Liquid outflow
QL
QG
QGR
Eq. 4.6
Where, C is the concentration (M), N is the amount
of methane in the expanded headspace of the serum bottle
(mol), and Vsample is the volume of the sample (L).
4.7.3.4 Calculations
V = Vs +VHS
n
Vsample
VG,1
R
QL(t)
Subsystem 1:
Liquid + gas bubbles
RVR
RV
CRL (t)
CL(t)
QGR(t)QGin,R = 0
Aeration reactor
QLR(t) C
in,R
L
=0
Liquid inflow reactor
VL
Stripping flask
Scum trap flask
Figure 4.5 Lay-out of the reactor (left) and the gas stripping device (right) for monitoring dissolved gases (Mampaey et al., 2015).
A liquid sample stream from the reactor is
continuously supplied to the stripping flask at a constant
flow rate QL, while maintaining a constant liquid volume
VL in the stripping flask. The scum trap flask is an empty
bottle to collect entrained scum from the stripping flask.
Nitrogen is used as the stripping gas in the stripping flask
through fine bubble aeration at a constant flow rate QGin.
The gas outflow of the gas-stripping device is analysed
by an online gas-phase analyser. The dissolved
OFF-GAS EMISSIONS TESTS
197
concentration in the reactor, CLR(t), is calculated from the
measured gas concentration CG,2(t) according to Eq. 4.7.
CRL (t ) =
QG
QL
Q
QL
⋅ CG,2 (t ) − G ⋅ a1
⋅ 1 +
⋅
a
V
QL
3
L
Eq. 4.7
The parameters a1 and a3 are determined from a
batch-stripping test. During this test, the stripping flask is
filled in batches with a liquid sample from the reactor
under study from which the dissolved N2O is
subsequently stripped with N2. The monitored gas phase
profile CG,2(t) from the stripping device is then described
by Eq. 4.8.
The gas-stripping device provides an adequate
method to indirectly measure dissolved gases (N2O or
other gases) in the liquid phase, for aerated as well as
non-aerated conditions/reactors, following variations
both in time and in space. Its application to an
intermittently aerated (on/off) partial nitritation
(SHARON) reactor was demonstrated by Mampaey et al.
(2015). Castro-Barros et al. (2015) applied the method to
a one-stage partial nitritation-anammox reactor, subject
to alternating high and low aeration. In both cases, the
mass balance approach on which the liquid N2O
concentration measurement method is based also allowed
the determination of the N2O formation rate.
CG,2 ( t ) = a1 + a 2 ⋅ exp ( −a3 ⋅ t ) − a 4 ⋅ exp ( −a5 ⋅ t ) Eq. 4.8
Liquid discharge pump
QG CGin
Stripping gas
Measured gas
concentration
CG,2 (t)
Liquid
sampling pump
QL
Unknown liquid
concentration
CLR (t)
VL
Stripping
flask
Scum
trap flask
Condensing
unit
Figure 4.6 Layout of the gas-stripping device.
4.7.4.2 Equipment
The detailed layout of the gas-stripping device is shown
in Figure 4.6, and consists of:
a. A stripping flask: a graded plastic cylinder of 250 mL
with an (aquarium) aeration stone for the stripping
gas and a rubber stopper with four connections
(stripping gas in, stripping gas out, sampled liquid in
and liquid out) to seal the stripping flask. The liquid
is introduced at the bottom of the stripping flask, next
to the aeration stone for intense mixing. The liquid is
extracted 10 cm below the rubber stopper, which
results in a constant liquid volume VL of 100 mL. A
larger liquid volume increases the sensitivity of the
device (K) but decreases the frequency of
measurements (or decreases the observability of rapid
changes in concentrations).
b. A scum trap flask: a 2 L bottle to collect any entrained
scum was used.
c. A liquid sampling pump: a peristaltic pump with a
tube with an inside diameter of 3.1 mm (Masterflex
tubing size L/S = 16), which yields a liquid flow rate
of 85 mL min-1. It is advised to keep the length of the
liquid feed tube as short as possible. A larger liquid
flow rate (QL) increases the sensitivity of the device.
198
The gas phase is measured by a gas-phase
measurement device as described in Section 4.5.
4.7.4.3 Calibration batch test
The parameters a1 and a3 are characteristics of the
gas-stripping device and are determined from a batchstripping test. During this test, the stripping flask is filled
in batches with a liquid sample (containing dissolved
N2O) from the reactor under study and flushed with the
stripping gas (typically N2). The monitored gas-phase
profile CG,2(t) from the stripping device is then described
by a double exponential profile (Eq. 4.9):
CG,2 ( t ) = a1 + a 2 ⋅ exp ( −a3 ⋅ t ) − a 4 ⋅ exp ( −a5 ⋅ t ) Eq. 4.9
Eq. 4.9 is fitted to the batch test measurement to
obtain the values for the coefficients. A calibration
example is shown in Figure 4.7.
Parameter a1 is related to the (N2O) conversion rate in
the stripping flask and its concentration in the stripping
gas, a3, is related to the interphase transfer rate in the
stripping flask, and a5 is related to the gas delay by the
scum trap flask. The exact meanings of these parameters
can be found in Mampaey et al. (2015).
27
Related to a5
Stripping gas N2O (ppm)
d. A liquid discharge pump: a peristaltic pump with a
tube with an inside diameter of 6.4 mm (Masterflex
tubing size L/S = 17), which yields a liquid flow rate
of 245 mL min-1.
e. Stripping gas: a stripping gas flow rate (QG) of 1.2
NL min-1 is required for the analyser. The following
two possibilities of stripping gas can be utilized: (i)
gas cylinder N2 is utilized, a mass flow controller can
be used to control the stripping gas flow rate (at 1.2
NL min-1), and (ii) ambient air can also be utilized as
stripping gas. A diaphragm pump in combination
with a mass flow controller can be used to provide the
required gas flow rate (1.2 NL min-1).
f. A condensing column / gas drying unit for the
removal of air moisture.
g. Tubing: it is advised to use neoprene tubing to
prevent diffusion of O2.
h. A weatherproof box to protect the electronic part of
the setup.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
18
Related to a3
9
Data fit
Measured stripping gas N2O concentration
0
0
2
4
6
8
10
12
Time (min)
Figure 4.7 Example of a calibration batch test with data fit.
4.7.4.4 Measurement accuracy
The sensitivity of the setup can be calculated from
Eq. 4.6. In the current setup, an accuracy for CLR(t) of
0.03 g N m-3 is achieved. The fastest measurable changes
in (CLR(t)) is estimated from Eq. 4.11, given as a
frequency of sampling (fsample).
K=
a3
Q
⋅ L
a 3 + D L QG
fsample < a 3 + D L
Eq. 4.10
Eq. 4.11
The parameter a3 is the decreasing exponential term
in the calibration batch test (Eq. 4.8), DL = QL/VL is the
liquid dilution rate, QG is the stripping gas flow rate (1.2
NL min-1), QL is the liquid sampling flow rate (84 mL
min-1), and VL is the liquid volume in the stripping flask
(100 mL).
4.7.4.5 Calculation of the N2O formation rate in the
stripping device
N2O formation can occur in the stripping flask and is
reflected in parameter a1. The N2O formation rate in the
gas-stripping device is calculated from Eq. 4.12. CGin is
the N2O concentration in the fresh stripping gas; if no
N2O is present, CGin = 0.
(
)
R V = a1 − Cin
G ⋅
QG
VL
Eq. 4.12
OFF-GAS EMISSIONS TESTS
199
The formation rate in the sampled water body is
calculated as the sum of the gas emission and change in
dissolved N2O through a mass balance, Eq. 4.13.
RV (t) =
ΔCR
L (t)
Δt
+ Emission
Eq. 4.13
4.8 DATA ANALYSIS AND PROCESSING
4.8.1 Determination of fluxes
The net flux of gaseous N2O or CH4 (kg m-2 d-1) can be
calculated based on the gas flow rate out of the flux
chamber (Qemission, m3 d-1), the gas concentration (C, kg
m-3), and the cross-sectional area of the SEIFC (A, m2)
(Eq. 4.14).
zone in the facility to the daily influent total Kjeldahl
nitrogen (TKN) loading according to Eq. 4.15.
Correspondingly, for methane, the measured flux could
be normalized to the influent COD loading. Emission
fractions are typically averaged over the course of the
diurnal sampling period and reported as the average (avg)
± standard deviation (sd) for each individual process
sampled.
During each campaign, wastewater nitrogen species
and COD concentrations including influent, bioreactor,
and effluent concentrations are measured simultaneously
about six times per day according to APHA et al. (2012)
to supplement the gas-phase measurements. The discrete
measurements are to be averaged to generate the
emission fractions described in Eq. 4.15.
Emission fraction =
Flux =
Qemission · C
A
Eq. 4.14
The calculated flux should be corrected to reflect
standard temperature (20 °C) and pressure (1 atm).
4.8.2 Determination of aggregated
emission fractions
As previously described, the specific locations selected
for measuring gaseous and aqueous concentrations of
N2O and CH4 can be close to the influent or effluent end
of each demarcated anoxic or aerobic zone in the WWTP,
or, alternatively, at locations where the production of
these two gases could be inferred based on initial
screening of the process variable concentrations. Using
these point measurements and assuming that these
measured concentrations were uniform over a particular
zone, the emissions from any given zone can be
computed according to the following equation:
Emissions from ith zone in a WWTP = Emissions from
SEIFC × (Area of ith zone/Area of SEIFC).
Emission measurements must be repeated at multiple
zones for any given plant. The overall emissions of the
plant are then calculated by summing the emissions from
each zone measured over a 24-h period (at the very
minimum). The surface flux calculated from Eq. 4.14 is
translated into the flux of a given zone by multiplying
with the specific zone area. The N2O emission fractions
(mass/mass) for each WWTP at any given time point can
be computed by normalizing the measured flux from each
∑ni=1 Fluxi × Areai (kg N2 O-N)
Daily influent TKN load (kg-N)
Eq. 4.15
Where, Fluxi is the N2O emission flux calculated
from the ith zone (kg N2O-N m-2 d-1), Areai is the surface
area of the ith zone (m2), n is the number of zones in a
given facility from which the N2O fluxes are captured,
and the Daily influent TKN load is the average influent
load (influent flow rate × influent TKN concentrations)
over the course of 24 h. It should be noted that the above
calculations reflect the emission factor calculated from
discrete N2O measurements. In plants where significant
diurnal variability exists, such variability will be
accounted for by a combination of explicit measurements
in select zones and mathematical modeling output of N2O
fluxes from the remaining zones.
Virtually identical equations can be used for methane,
with the only distinction of normalizing to total COD
mass loads in the influent or, as applicable, the COD
mass removal rate.
On average, wastewater characterization about six
times per day is recommended at each gas sampling
location as well as in the tank influent and effluent. At
facilities where analysis is not as frequent (for instance in
the influent and effluent samples), daily composite
measurements can be used. Alternatively, in some
facilities, online devices (for measuring pH, DO,
oxidation-reduction potential (ORP), and select Nspecies, including NH4+-N and NO3--N) can also be used
at different locations of the activated sludge tank to
facilitate the characterization of the wastewater.
200
4.8.3 Calculation of the emission factors
To directly compare results from off-gas monitoring
campaigns with the approaches suggested by the EPA
and IPCC, it becomes necessary to summarize the
monitoring campaign results in terms of emission factors.
Such emission factors can be computed by normalizing
the total reactor N2O or CH4 mass flux to the unit
population equivalent flow rate (100 gal PE-1 d-1, or
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
equivalently, 378.5 L PE-1 d-1in the US, US EPA, 2012).
The resulting emissions fluxes are expressed in units
consistent with the US EPA inventory report (g N2O or
CH4 PE-1 d-1) (US EPA, 2012). Although the dynamic
variability in WWTP off-gas emissions renders the use of
point emission factors (such as the ones presented by the
EPA and IPCC) somewhat limited, the added step of
computing emission factors could be useful for a side-byside comparison of estimated emissions and those that are
actually measured.
References
Ahn, J.-H., Kim, S., Pagilla, K., Katehis, D., and Chandran, K.
(2010a). Spatial and temporal variability in N2O generation and
emission from full-scale BNR and non-BNR processes. Water
Environment Research. 82: 2362-2372.
Ahn, J. H., Kim, S., Park, H., Rahm, B., Pagilla, K., and Chandran,
K. (2010b). N2O Emissions from Activated Sludge Processes,
2008-2009: Results of a National Monitoring Survey in the
United States. Environmental Science & Technology. 44: 45054511.
APHA, AWWA and WEF, (2005). Standard methods for the
examination of water and wastewater. 22nd edition. Eaton, A.D.,
Clesceri, L.S., and Greenberg, A.E. (eds.) Washington DC.
Castro-Barros C.M., Daelman, M.R.J., Mampaey, K.E., van
Loosdrecht M.C.M., Volcke, E.I.P. (2015). Effect of aeration
regime on N2O emission from partial nitritation-anammox in a
full-scale granular sludge reactor, Water Res. 68: 793-803.
Daelman, M.R.J., van Voorthuizen E.M., van Dongen, U.G.J.M.,
Volcke, E.I.P., and van Loosdrecht M.C.M., (2012). Methane
emission during municipal wastewater treatment. Water Res.
46: 3657-3670.
Daelman, M.R., de Baets, B., van Loosdrecht, M.C.M., and Volcke,
E.I. (2013). Influence of sampling strategies on the estimated
nitrous oxide emission from wastewater treatment
plants. Water Res. 47(9): 3120-3130.
Daelman, M., van Eynde, T., van Loosdrecht, M.C.M., Volcke,
E.I.P. (2014) Effect of Process design and operating parameters
on aerobic methane oxidation in municipal WWTP. Water Res.
66, 308-319.
Foley, J., de Haas, D., Yuan, Z., and Lant, P. (2010). Nitrous oxide
generation in full-scale biological nutrient removal wastewater
treatment plants. Water Res. 44(3): 831-844.
Gal'chenko, V.F., Lein, A.Y., and Ivanov, M.V. (2004). Methane
content in the bottom sediments and water column of the Black
Sea. Microbiology. 73(2): 211-223.
Gapes, D., Pratt, S., Yuan, Z., and Keller, J. (2003). Online
titrimetric and off-gas analysis for examining nitrification
processes in wastewater treatment. Water Res. 37(11): 26782690.
Hellinga, C., Vanrolleghem, P., Van Loosdrecht, M.C.M., and
Heijnen, J.J. (1996). The potential of off-gas analyses for
monitoring wastewater treatment plants. Water Sci Tech. 33(1):
13-23.
IPCC (2013). Climate Change 2013: The Physical Science Basis.
Working Group I Contribution to the Fifth Assessment Report
of the Intergovernmental Panel on Climate Change. Chapter 8:
Anthropogenic and Natural Radiative Forcing.
Kampschreur, M.J., Tan, N.C.G., Kleerebezem, R., Picioreanu, C.,
Jetten M.S.M., and van Loosdrecht M.C.M. (2008a). Effect of
Dynamic Process Conditions on Nitrogen Oxides Emission
from a Nitrifying Culture. Environmental Science and
Technology. 42: 429-435.
Kampschreur, M.J., Temmink H., Kleerebezem, R., Jetten, M.S.M.,
and van Loosdrecht M.C.M. (2009). Nitrous oxide emission
during wastewater treatment. Water Res. 43: 4093-4103.
Kampschreur, M. J., van der Star W.R.L., Wielders H.A., Mulder
J.W., Jetten M.S.M., and van Loosdrecht M.C.M. (2008b).
Dynamics of nitric oxide and nitrous oxide emission during
full-scale reject water treatment. Water Res. 42: 812-826.
Kienbusch, M. (1986). Measurement of Gaseous Emissions Rates
from Land Surfaces using an Emission Isolation Flux Chamber,
User's Guide, EPA Users Guide. United States Environmental
Protection Agency.
Klopfenstein Jr,R. (1998). Air velocity and flow measurement using
a Pitot tube. ISA transactions. 37(4): 257-263.
Mampaey, K. E., van Dongen, U. G., van Loosdrecht, M.C.M.,
Volcke, E.I. (2015). Novel method for online monitoring of
dissolved N2O concentrations through a gas stripping
device. Environmental Technology. 36(13): 1680-1690.
Melcer, H. (1999). Methods for Wastewater Characterization in
Activated Sludge Modelling. Water Environment Research
Foundation, Alexandria, VA.
Puig, S., van Loosdrecht, M.C.M., Colprim, J., Meijer, S.C.F.
(2008). Data evaluation of full-scale wastewater treatment
plants by mass balance. Water Res. 42(18): 4645-4655.
Pratt, S., Yuan, Z., Gapes, D., Dorigo, M., Zeng, R.J., and Keller, J.
(2003). Development of a novel titration and off‐gas analysis
(TOGA) sensor for study of biological processes in wastewater
treatment systems. Biotechnology and Bioengineering. 81(4):
482-495.
Rapson, T.D., Dacres, H. (2014). Analytical techniques for
measuring nitrous oxide. TrAC Trends in Analytical
Chemistry. 54: 65-74.
Rieger, L., Gillot, S., Langergraber, G., Ohtsuki, T., Shaw, A., Tak,
I., and Winkler, S. (2012). Guidelines for using activated sludge
models. Water Intelligence Online, 11, 9781780401164.
Tata, P., Witherspoon J., and Lue-Hing C. (Eds). (2003). VOC
Emissions from Wastewater Treatment Plants. Lewis
Publishers, Boca Raton, FL.
USEPA. 2012. Inventory of U.S. Greenhouse Gas Emissions and
Sinks: 1990-2006, EPA 430-R-08-005. Washington, D.C.
Weiss, R.F. (1981). Determinations of carbon dioxide and methane
by dual catalyst flame ionization chromatography and nitrous
oxide by electron capture chromatography. Journal of
Chromatographic Science. 19(12): 611-616.
5
DATA HANDLING AND PARAMETER ESTIMATION
Authors:
Reviewer:
Gürkan Sin
Krist V. Gernaey
Sebastiaan C.F. Meijer
Juan A. Baeza
5.1 INTRODUCTION
Modelling is one of the key tools at the disposal of
modern wastewater treatment professionals, researchers
and engineers. It enables them to study and understand
complex phenomena underlying the physical, chemical
and biological performance of wastewater treatment
plants at different temporal and spatial scales.
At full-scale wastewater treatment plants (WWTPs),
mechanistic modelling using the ASM framework and
concept (e.g. Henze et al., 2000) has become an
important part of the engineering toolbox for process
engineers. It supports plant design, operation,
optimization and control applications. Models have also
been increasingly used to help take decisions on complex
problems including the process/technology selection for
retrofitting, as well as validation of control and
optimization strategies (Gernaey et al., 2014; MauricioIglesias et al., 2014; Vangsgaard et al., 2014; Bozkurt et
al., 2015).
Models have also been used as an integral part of the
comprehensive analysis and interpretation of data
obtained from a range of experimental methods from the
laboratory, as well as pilot-scale studies to characterise
and study wastewater treatment plants. In this regard,
models help to properly explain various kinetic
parameters for different microbial groups and their
activities in WWTPs by using parameter estimation
techniques. Indeed, estimating parameters is an integral
part of model development and application (Seber and
Wild, 1989; Ljung, 1999; Dochain and Vanrolleghem,
2001; Omlin and Reichert, 1999; Brun et al., 2002; Sin
et al., 2010) and can be broadly defined as follows:
Given a model and a set of data/measurements from
the experimental setup in question, estimate all or some
of the parameters of the model using an appropriate
statistical method.
The focus of this chapter is to provide a set of tools
and the techniques necessary to estimate the kinetic and
stoichiometric parameters for wastewater treatment
processes using data obtained from experimental batch
activity tests. These methods and tools are mainly
© 2016 Gürkan Sin and Krist V. Gernaey. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen. C.M. Lopez-Vazquez and D. Brdjanovic. ISBN:
9781780404745 (Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
202
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
intended for practical applications, i.e. by consultants,
engineers, and professionals. However, it is also
expected that they will be useful both for graduate
teaching as well as a stepping stone for academic
researchers who wish to expand their theoretical interest
in the subject. For the models selected to interpret the
experimental data, this chapter uses available models
from literature that are mostly based on the Activated
Sludge Model (ASM) framework and their appropriate
extensions (Henze et al., 2000).
The chapter presents an overview of the most
commonly used methods in the estimation of parameters
from experimental batch data, namely: (i) data handling
and validation, (ii) parameter estimation: maximum
likelihood estimation (MLE) and bootstrap methods, (iii)
uncertainty analysis: linear error propagation and the
Monte Carlo method, and (iv) sensitivity and
identifiability analysis.
5.2 THEORY AND METHODS
5.2.1 Data handling and validation
5.2.1.1 Systematic data analysis for biological
processes
Most activated sludge processes can be studied using
simplified process stoichiometry models which rely on a
‘black box’ description of the cellular metabolism using
measurement data of the concentrations of reactants
(pollutants) and products e.g. CO2, intermediate oxidised
nitrogen species, etc. Likewise, the Activated Sludge
Model (ASM) framework (Henze et al., 2000) relies on
a black box description of aerobic and anoxic
heterotrophic activities, nitrification, hydrolysis and
decay processes.
A general model formulation of the process
stoichiometry describing the conversion of substrates to
biomass and metabolic products is formulated below (for
carbon metabolism):
CH 2O + YSO O 2 + YSN NH 3
→ YSX X + YSCCO 2 + YSP1P1 …+ YSW H 2O
Eq. 5.1
Equation 5.1 represents a simplification of the
complex metabolic ‘machinery’ of cellular activity into
one global relation. This simplified reaction allows the
calculation of the process yields including YSO (yield of
oxygen per unit substrate), YSN (yield of nitrogen per unit
substrate), YSX (yield of biomass per unit substrate), YSC
(yield of CO2 per unit of substrate), YSP1 (yield of
intermediate product P1 per unit of substrate), and YSW
(yield of water per unit of substrate).
The coefficients of this equation are written on the
basis of 1 C-mol of carbon substrate. This includes
growth yield for biomass, YSX, substrate (ammonia)
consumption yields, YSN, oxygen consumption yields,
YSO, yield for production of CO2, YSC, and yield for
water, YSW. The biomass, X, is also written on the basis
of 1 C-mol and is assumed to have a typical composition
of CHaObNc. The biomass composition can be measured
experimentally, CH1.8O0.5N0.2 being a typical value.
Some of the yields are also measured experimentally
from the observed rates of consumption and production
of components in the process as follows:
Yji =
ri q i
=
rj q j
and Yji = Yij−1
Eq. 5.2
Where,
qi
refers
to
the
volumetric
conversion/production rate of component i, i.e. the mass
of component i per unit volume of the reactor per unit
time (Mass i Volume-1 Time-1), ri refers to the measured
rate of the mass of component i per unit time per unit
weight of the biomass (Mass i Time-1 Mass biomass-1)
and Yji is the yield of component i per unit of component
j. In the case of biomass, x, this would refer to the
specific growth rate μ:
μ = rx =
qx
x
Eq. 5.3
One of the advantages of using this process
stoichiometry is that it allows elemental balances for C,
H, N and O to be set up and to make sure that the process
stoichiometry is balanced. For the process stoichiometry
given in Eq 5.1, the following elemental balance for
carbon will hold, assuming all the relevant yields are
measured:
C − balance : − 1 + Ysx + Ysc + Ysp1 = 0
Eq. 5.4
Similarly to the carbon balance, the elemental
balance for N, O and H can also be performed. Usually
in biological process studies, the yield coefficient for
water, YSW, is ignored because the production of water is
negligible compared with the high flow rates typically
treated in WWTPs. For this reason, H and O balances and
DATA HANDLING AND PARAMETER ESTIMATION
process stoichiometry are usually not closed in
wastewater applications. However, the balance for the
degree of reduction is closed in wastewater treatment
process stoichiometry. This is the framework on which
ASM is based. The degree of reduction balance is
relevant since most biological reactions involve
reduction-oxidation (redox)-type chemical conversion
reactions in metabolism activities.
5.2.1.2 Degree of reduction analysis
A biological process will convert a substrate i.e. the input
to a metabolic pathway, into a product that is in a reduced
or oxidized state relative to the substrate. In order to
perform redox analysis on a biological process, a method
to calculate the redox potential of substrates and products
is required. In the ASM framework and other
biotechnological applications (Heijnen, 1999; Villadsen
et al., 2011), the following methodology is used:
1) Define a standard for the redox state for the balanced
elements, typically C, O, N, S and P.
2) Select H2O, CO2, NH3, H2SO4, and H3PO4 as the
reference redox-neutral compounds for calculating
the redox state for the elements O, C, N, S, and P
respectively. Moreover, a unit of redox is defined as
H = 1. With these definitions, the following redox
levels of the five listed elements are obtained: O = -2,
C = 4, N = -3, S = 6 and P = 5.
3) Calculate the redox level of the substrate and
products using the standard redox levels of the
elements. Several examples are provided below:
a) Glucose (C6H12O6): 6 · 4 + 12 · 1 + 6 · (-2) = 24.
Per 1 C-mol, the redox level of glucose becomes,
γg = 24/6 = 4 mol e- C-mol-1.
b) Acetic acid (C2H4O2): 2 · 4 + 4 · 1 + 2 · (-2) = 8.
Per 1 C-mol, the redox level of Hac becomes,
γa = 8/2 = 4 mol e- C-mol-1
c) Propionic acid (C3H6O2): 3 · 4 + 6 · 1 + 2 · (-2) =
14. Per 1 C-mol, the redox level of HPr becomes,
γp = 14/3 = 4.67 mol e- C-mol-1.
d) Ethanol (C2H6O): 2 · 4 + 6 · 1 + 1 · (-2) = 12. Per
1 C-mol, the redox level of HAc becomes, γe =
12/2 = 6 mol e- C-mol-1.
4) Perform a degree of reduction balance over a given
process stoichiometry (see Example 5.1).
Example 5.1 Elemental balance and degree of reduction analysis
for aerobic glucose oxidation
General process stoichiometry for the aerobic oxidation
of glucose to biomass:
203
CH 2 O + YSO O 2 + YSN NH 3 → YSX X + YSC CO 2
Eq. 5.5
Assuming the biomass composition X is
CH1.8O0.5N0.2. The degree of reduction for biomass is
calculated assuming the nitrogen source is ammonia
(hence the nitrogen oxidation state is -3, γX: 4 + 1.8 + 0.5
· (-2) + 0.2 · (-3) = 4.2 mol e- C-mol-1.
Now C, N and the degree of reduction balances can
be performed for the process stoichiometry as follows:
Carbon balance: −1 + YSX + YSC = 0
Eq. 5.6
Nitrogen balance: −YSN + 0.2 ⋅ YSX = 0
Eq. 5.7
Redox balance:
−1 ⋅ γ g − γ O2 ⋅ YSO − γ NH3 ⋅ YSN + γ X ⋅ YSX + γ CO2 ⋅ YSC = 0
−1 ⋅ γ g − γ O2 ⋅ YSO − 0 ⋅ YSN + γ X ⋅ YSX + 0 ⋅ YSC = 0
Eq. 5.8.
In these balance equations, there are four unknowns
(YSN, YSO, YSX, YSC). Since three equations are available,
only one measurement of the yield is necessary to
calculate all the others. For example, in ASM
applications, biomass growth yield is usually assumed
measured or known, hence the other remaining yields
can be calculated as follows:
CO2 yield: YSC = 1 - YSX
Eq. 5.9
NH3 yield: YSN = 0.2YSX
Eq. 5.10
O2 yield: YSO =
γ g − γ x ⋅ YSX
γ O2
=
4 − 4.2YSX
4
Eq. 5.11
With these coefficients known, the process
stoichiometry model for 1 C-mol of glucose
consumption becomes as follows:
4 − 4.2YSX
⋅ O2 + 0.2YSX ⋅ NH 3 →
4
YSX ⋅ X + (1 − YSX ) ⋅ CO2
CH 2 O +
Eq. 5.12
In the ASM framework, the process stoichiometry is
calculated using a unit production of biomass as a
reference. Hence, the coefficients of Eq. 5.12 can be rearranged as follows:
204
4 − 4.2YSX
1
⋅ CH 2 O +
⋅ O2 + 0.2NH 3 →
YSX
4YSX
1
− YSX ⋅ CO2
X+
Y
SX
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Eq. 5.13
The unit conversion from a C-mol to a g COD basis,
being the unit of ASM models, is defined using O2 as the
reference compound. Accordingly, 1 g COD is defined
as -1 g O2. From the degree of reduction of oxygen, the
conversion to COD from one unit redox (mol e-) is
calculated as follows:
MWO2
Molecular weight of O2
=
=
Degree of reduction of O2
γO2
In addition to the elemental balances, the degree of
reduction balance provides information about whether
the right compounds are included for a given pathway or
whether a compound is missing in the process
stoichiometry. Adding this check is helpful and provides
consistency with the bioenergetic principles of biological
processes (Roels, 1980; Heijnen, 1999; Villadsen et al.,
2011).
The consistency checks and the elemental balances
(in addition to the charge balances) are included in the
ASM framework as a conservation matrix to verify the
internal consistency of the yield coefficients (Henze et
al., 2000).
32
= 8 g COD L−1 (mole− )−1
4
The elemental composition and degree of reduction
can be performed systematically using the following
generic balance equation in order to test the consistency
of the measured data:
To convert a C-mol to a g COD basis, the unit redox
needs to be multiplied with the degree of reduction of the
substrate as follows:
g COD
mol e g COD
= γg ⋅ 8
⋅
C − mol
C − mole mol e
Eq. 5.14
Eq. 5.15
5.2.1.3 Consistency check of the experimental data
The value of performing elemental balances around data
collected from experiments with biological processes is
obvious: to confirm the data consistency with the first
law of thermodynamics, which asserts that energy (in the
form of matter, heat, etc.) is conserved. A primary and
obvious requirement for performing elemental balances
is that the model is checked and consistent. Experimental
data needs to be checked for gross (measurement) errors
that may be caused by incorrect calibration or
malfunction of the instruments, equipment and/or
sensors.
Inconsistency in the data can be checked from the
sum of the elements that make up the substrates
consumed in the reaction (e.g. glucose, ammonia,
oxygen, etc.). This should equal to the sum of the
elements (products) produced in the reaction (therefore
also see Eq. 5.4 for the carbon balance). Deviation from
this elemental balance indicates an incorrectly defined
system description, a model inconsistency and/or
measurement flaws.
N
e q + exq x + j=1 epjqpj = 0
j=1 sj sj
M
Eq. 5.16
The equation above is formulated for a biological
process with N substrates and M metabolic products. In
the equation, e is the elemental composition (C, H, O and
N) for a component, and q the volumetric production (or
consumption) rate for substrates (qsj), biomass (qx) and
metabolic products (qpj). Hence, the elemental balance
can be formulated as follows:
E⋅q = 0
Eq. 5.17
In this equation, E is the conservation matrix and its
columns refer to each conserved element and property,
e.g. C, H, O, N, γ, etc. Each row of matrix E contains
values of a conserved property related to substrates,
products and biomass; q is a column vector including the
measured volumetric rates for each compound. This is
substrate as well as products and biomass.
The total number of columns in E is the number of
compounds, which is the sum of substrates (N), products
(M) and biomass, hence N + M + 1. The total number of
constraints is 5 (C, H, O, N and γ). This means that N+M4 is the number of degrees of freedom that needs to be
measured or specified in order to calculate all the rates.
Typically, not all the rates will be measured in batch
experiments. Therefore, let us assume qm is the measured
set of volumetric rates and qu the unmeasured set of rates
DATA HANDLING AND PARAMETER ESTIMATION
205
which need to be calculated. In this case Eq. 5.17 can be
reformulated as follows:
E m q m + E u q u =0
Eq. 5.18
q u = − ( E u ) E m q m =0
-1
Provided that the inverse of Eu exists (det(Eu) ≠ 0),
Eq. 5.18 provides a calculation/estimation of the
unmeasured rates in a biological process. These
estimated rates are valuable on their own, but can also be
used for validation purposes if redundant measurements
are available. This systematic method of data consistency
check is highlighted in Example 5.2.
All these calculations help to verify and validate the
experimental data and measurement of the process yield.
The data can now be used for further kinetic analysis and
parameter estimation.
5.2.2 Parameter estimation
Here we recall a state-space model formalism to describe
a system of interest. Let y be a vector of outputs resulting
from a dynamic model, f, employing a parameter vector,
θ; input vector, u; and state variables, x:
dx
= f ( x,θ, u, t ) ; x ( 0 ) = x 0
dt
y = g ( x,θ, u, t )
Eq. 5.19
The above equation describes a system (a batch setup
or a full WWTP) in terms of a coupled ordinary
differential equations (ODE) and algebraic system of
equations using a state-space formalism.
The problem statement for parameter estimation
reads then as follows: for a given set of measurements, y,
with its measurement noise collected from the system of
interest, and given the model structure in Eq. 5.19,
estimate the unknown model parameters (θ).
The solution approaches to this problem can be
broadly classified as the manual trial and error method,
and formal statistical methods.
5.2.2.1 The manual trial and error method
This approach has no formal scientific basis except for a
practical motivation that has to do with getting a good
model fit to the data. It works as follows: the user
chooses one parameter from the parameter set and then
changes it incrementally (increases or decreases around
its nominal value) until a reasonable model fit is obtained
to the measured data. The same process may be iterated
for another parameter. The fitting process is terminated
when the user deems that the model fit to data is good.
This is often determined by practical and/or time
constraints because this procedure will never lead to an
optimal fit of the model to the measured data. In addition,
multiple different sets of parameter values can be
obtained which may not necessarily have a physical
meaning. The success of this procedure often relies on
the experience of the modeller in selecting the
appropriate parameters to fit certain aspects of the
measured data. Although this approach is largely
subjective and suboptimal, the approach is still widely
used in industry as well as in the academic/research
environment. Practical data quality issues do not often
allow the precise determination of parameters. Also not
all (commercial) modelling software platforms provide
the appropriate statistical routines for parameter
estimation. There are automated procedures for model
calibration using algorithms such as statistical sampling
techniques, optimization algorithm, etc. (Sin et al.,
2008). However, such procedures focus on obtaining a
good fit to experimental data and not necessarily on the
identifiability and/or estimation of a parameter from a
data set. This is because the latter requires proper use of
statistical theory.
5.2.2.2 Formal statistical methods
In this approach, a proper statistical framework is used
to suggest the problem, which is then solved
mathematically by using appropriate numerical solution
strategies, e.g. minimization algorithms or sampling
algorithms. Under this category, the following statistical
frameworks are usually employed:
a. Frequentist framework (maximum likelihood, least
squares, non-linear regression, etc.).
b. Bayesian framework (Metropolis-Hasting, Markov
Chain Monte Carlo (MCMC), importance sampling,
etc.).
c. Pragmatic/hybrid framework (employing some
elements of the two schools of thought above, e.g. the
bootstrap method, Monte Carlo filtering, etc.).
The above statistical methods are among the most
commonly used and recommended here as well. In
particular, we focus on the frequentist and bootstrap
methods as they are more fit to the intended purpose of
this chapter.
206
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Frequentist method - maximum likelihood theory
In the parameter estimation problem we usually define
parameter estimators, θ , to distinguish them from the
true model parameters, θ. In the context of statistical
estimation, model parameters are defined as unknown
and statistical methods are used to infer their true value.
This difference is subtle but important to understand and
to interpret the results of parameter estimation,
irrespective of the methods used.
Maximum likelihood is a general method for finding
estimators, θ, from a given set of measurements, y. In
this approach, the model parameters θ are treated as true,
fixed values, but their corresponding estimators θ are
treated as random variables. The reason is that the
estimators depend on the measurements, which are
assumed to be a stochastic process:
y = f (θ) + ε
where
ε ∝ N ( 0, σ )
Eq. 5.20
Measurement errors, ε, are defined by a probability
distribution, e.g. normal distribution, N, with zero mean
and standard deviation (σ). With these assumptions, the
likelihood function (L) for the parameter estimation
becomes as follows (Seber and Wild, 1989):
L ( y,θ ) =
( y − f ( θ ) )2
1
exp −
2σ 2
σ 2π
S ( y,θ ) =
( y − f (θ))
2
σ2
Eq. 5.23
Where, y stands for the measurement set, (θ) stands
for the corresponding model predictions, and Σ stands for
the standard deviation of the measurement errors. The
solution to the objective function (Eq. 5.24) is found by
minimization algorithms (e.g. Newton’s method,
gradient descent, interior-point, Nelder-Mead simplex,
genetic, etc.).
θ̂: min θ S ( y,θ )
∂
S ( y,θ ) =0
∂θ
θ̂
Eq. 5.24
The solution to the above optimization problem
provides the best estimate of the parameter values. The
next step is to evaluate the quality of the parameter
estimators. This step requires the estimation of the
confidence interval of the parameter values and the
pairwise linear correlation between the parameters.
The covariance matrix of parameter estimators
Eq. 5.21
The most likely estimate of θ is found as those
parameter values that maximize the likelihood function:
θ̂: min θ L ( y,θ )
equivalent to minimizing the following cost (or
objective) function, S(y,θ) (Seber and Wild, 1989):
Eq. 5.22
The solution to this problem setting (5.24) is often
found by optimization algorithms such as simplex,
interior point, genetic algorithms, simulated annealing,
etc. The parameters obtained by calculating the
maximum likelihood (Eq. 5.21) are the same as the
parameters obtained by calculating the minimum cost
function in Eq. 5.23.
The least squares method
This is a special case of the maximum likelihood method
in which the measurements are assumed to be
independent and identically distributed with white
measurement errors having a known standard deviation,
σ (Gaussian). The likelihood function becomes
As a result of stochastic measurement, estimators have a
degree of uncertainty. In the frequentist framework of
thought, probability is defined in terms of the frequency
of the occurrence of outcomes. Hence, in this method the
uncertainty of the parameter estimators is defined by a
95 % confidence interval interpreted as the range in
which 95 times out of 100 the values of the parameter
estimators are likely to be located. This can be explained
as if one performs the same measurement 100 times, and
then performs the parameter estimation on these 100 sets
and observes the following: 95 occurrences of the
estimator values lie in the confidence interval, while 5
occurrences are outside this interval.
In order to estimate the confidence interval, first the
covariance matrix (cov θ ), which contains complete
information about the uncertainty of the parameter
estimators of the estimators, needs to be estimated. One
method to obtain cov θ is to use a linear approximation
method through estimation of the Jacobian matrix (F.) of
the parameter estimation problem (Seber and Wild,
1989):
DATA HANDLING AND PARAMETER ESTIMATION
()
−1
cov θˆ = s 2 ( F.'⋅ F.)
where F. =
207
∂f ( θ )
∂θ
θ =θ
Eq. 5.25
Where, s2 is the unbiased estimation of σ2 obtained
from the residuals of the parameter estimation:
s2 =
( )
Smin y,θˆ
Eq. 5.26
n−p
Here, n is the total number of measurements, p is the
number of estimated parameters, n-p is the degrees of
freedom, Smin (y,θ) is the minimum objective function
value and F. is the Jacobian matrix, which corresponds to
the first order derivative of the model function, f, with
respect to the parameter vector θ evaluated at θ = θ.
The covariance matrix is a square matrix with (p×p)
dimensions. The diagonal elements of the matrix are the
variance of the parameter estimators, while the nondiagonal elements are the covariance between any pair of
parameter estimators.
The 95 % confidence interval of the parameter
estimators can now be approximated. Assuming a large
n, the confidence intervals (the difference between the
estimators and true parameter values), follow a student tdistribution, the confidence interval at 100 (1-α) %
significance:
θ1-α = θ ± t α/2
N-p diag cov θ
()
Eq. 5.27
Where, tN-pα/2 is the upper α/2 percentile of the tdistribution with N-p degrees of freedom, and
diag cov(θ) represents the diagonal elements of the
covariance matrix of the parameters.
The pairwise linear correlation between the
parameter estimators, Rij, can be obtained by calculating
a correlation matrix from unit standardization of the
covariance matrix as follows:
R ij =
cov ( θ i ,θ j )
σ θi × σ θ j
Eq. 5.28
This linear correlation will range from [-1 1] and
indicate whether or not he parameter estimator is
uniquely identifiable (if the correlation coefficient is
low) or correlated (if the correlation coefficient is high).
The bootstrap method
One of the key assumptions for using the maximum
likelihood estimation (MLE) method as well as its
simplified version, the nonlinear least squares method, is
that the underlying distribution of errors is assumed to
follow a normal (Gaussian) distribution.
In many practical applications, however, this
condition is rarely satisfied. Hence, theoretically the
MLE method for parameter estimation cannot be applied
without compromising its assumptions, which may lead
to over or underestimation of the parameter estimation
errors and their covariance structure.
An alternative to this approach is the bootstrap
method developed by Efron (1979), which removes the
assumption that the residuals follow a normal
distribution. Instead, the bootstrap method works with
the actual distribution of the measurement errors, which
are then propagated to the parameter estimation errors by
using an appropriate Monte Carlo scheme (Figure 5.1).
The bootstrap method uses the original data set D(0)
with its N data points, to generate any number of
synthetic data sets DS(1);DS(2);…., also with N data
points. The procedure is simply to draw N data points
with replacements from the set D(0). Because of the
replacement, sets are obtained in which a random
fraction of the original measured points, typically 1/e =
37 %, are replaced by duplicated original points. This is
illustrated in Figure 5.1.
The application of the bootstrap method for
parameter estimation in the field of wastewater treatment
requires adjustment due to the nature of the data that is
in the time series. Hence, the sampling is not performed
from the original data points (which are the time series
and indicate a particular trend). Instead, the sampling is
performed from the residual errors and then added to the
simulated model outputs (obtained by using reference
parameter estimation) (Figure 5.1). This is reasonable
because the measurement errors are what is assumed to
be stochastic and not the main trend of the measured data
points,
which
are
caused
by
biological
processes/mechanisms. Bearing this in mind, the
theoretical background of the bootstrap method is
outlined below.
208
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
^
Synthetic data y1
^s
^
Monte Carlo Parameter θ1
^
Synthetic data y2
^s
^
Monte Carlo Parameter θ2
^
Synthetic data y3
^s
^
Monte Carlo Parameter θ3
^
Synthetic data y4
^s
^
Monte Carlo Parameter θ4
Random sampling from
^
residuals, e
Experimental
Data Set, y
Reference Parameter
^
Estimation, θ
Figure 5.1 Illustration of the workflow for the bootstrap method: synthetic data sets are generated by Monte Carlo samples (random sampling with
replacement) from the reference MLE. For each data set, the same estimation procedure is performed, giving M different sets of parameter estimates: θS(1),
θS(2), … θS(M).
Let us define a simple nonlinear model where yi is the
ith measurement, fi is the ith model prediction, θ is a
parameter vector (of length p), and εi is the measurement
error of yi:
yi = fi ( θ ) + ε i
where
εi ∞ F
Eq. 5.29
The distribution of errors, F, is not known. This is
unlike in MLE, where the distribution is assumed a
priori. Given y, use least squares minimization, to
estimate θ:
θ̂ : min θ y − f ( θ )
2
Eq. 5.30
The bootstrap method defines F as the sample
probability distribution of ε as follows:
1
F̂ =
(density) at εi = ( yi − fi ( θ ) )
n
i = 1,2,…n
Eq. 5.31
th
The density is the probability of the i observation.
In a uniform distribution each observation (in this case
the measurement error, εi) has an equal probability of
occurrence, where density is estimated from 1/n. The
bootstrap sample, y*, given (, F), is then generated as
follows:
()
y*i = f i θˆ + ε*i
where ε*i ∝ Fˆ
Eq. 5.32
The realisation of measurement error in each
bootstrap method, ε∗ , is simulated by random sampling
with replacement from the original residuals, which
assigns each point with a uniform (probability) weight.
By performing N random sampling with a replacement
and then adding them to the model prediction (Eq. 5.31),
a new synthetic data set is generated, Ds(1) = y*.
By repeating the above sampling procedure M times,
M data sets are generated: Ds(1), Ds(2), Ds(3), … Ds(M).
Each synthetic data set, Ds(j), makes it possible to
obtain a new parameter estimator θ(j) by the same least
squares minimisation method which is repeated M times:
θ̂ j : minθ Ds ( j) − f ( θ )
2
where
j = 1,2…M
Eq. 5.33
The outcome from this iteration is a matrix of
parameter estimators, θ(M × p) (M is the number of
Monte Carlo samples of synthetic data and p is the
number of parameters estimated). Hence, each parameter
estimator now has a column vector with values. This
vector of values can be plotted as a histogram and
interpreted using common frequentist parameters such as
the mean, standard deviation and the 95 % percentile.
The covariance and correlation matrix can be computed
using θ (M × p) itself. This effectively provides all the
needed information on the quality of the parameter
estimators.
DATA HANDLING AND PARAMETER ESTIMATION
209
For measurement errors that follow a normal
distribution, both MLE and the bootstrap method will
essentially provide the same results. However, if the
underlying distribution of the measurements
significantly deviates from a normal distribution, the
bootstrap method is expected to provide a better analysis
of the confidence interval of the estimators.
5.2.3 Uncertainty analysis
5.2.3.1 Linear error propagation
In linear error propagation, the covariance matrix of the
parameter estimators, cov(θ), is used to propagate
measurement errors to model prediction errors and to
calculate standard errors and confidence intervals of the
parameter estimates. Therefore, the covariance matrix of
model predictions, cov(y), can be estimated using
cov θ as follows (Seber and Wild, 1989):
()
−1
cov ( y ) = ( F.'⋅ F.) cov θˆ ( F.'⋅ F.)
Eq. 5.34
In a similar fashion, the 1-α confidence interval of the
predictions, y, can be approximated as follows:
y1−α = y ± t α/2
N-p diag cov ( y )
Eq. 5.35
This concludes parameter estimation, confidence
intervals and prediction uncertainty as viewed from the
point of view of the frequentist analysis.
I = f ( x ) dx = f ( u1…u d ) d d x
Eq. 5.36
Authors consider the integral of a function f(x) with
x as the input vector x = (u1,…ud). Hence, the integral is
taken on the d variables u1, .., ud over the unit hypercube
[0, 1]d. In the parameter estimation, these input variables
are parameters of the model that have a certain range
with lower and upper bounds. We assume that f is
square-integrable, which means that a real value solution
exists at each integration point. As a short-hand notation
we will denote a point in the unit hypercube by x = (u1, ..
ud) and the function evaluated at this point by f(x) = f(u1,
.. ud), and then the multidimensional integration
operation is given by:
E=
1 N
f (xN )
N 1
lim
N →∞
1 N
f (xN ) = I
N 1
Eq. 5.37
The law of large numbers ensures that the MC
estimate (E) converges to the true value of this integral.
However, as most of the time N is finite (a sampling
number from input space u with dxd dimension), there
will be an error in the Monte Carlo integration of
multidimensional functions. This Monte Carlo
integration error is scaled like 1/√N. Hence, the average
Monte Carlo integration error is given by
MCerr = (f)/√N, where σ(f) is the standard deviation of
the error, which can be approximated using sample
variance:
2
1 N
(f ( x N ) − E)
N −1 1
5.2.3.2 The Monte Carlo method
σ2 ( f ) ≈ s2 ( f ) =
The Monte Carlo (MC) method was originally used to
calculate multi-dimensional integrals and its systematic
use started in the 1940s with the ‘Los Alamos School’ of
mathematicians and physicists, namely Von Neumann,
Ulam, Metropolis, Kahn, Fermi and their collaborators.
The term was coined by Ulam in 1946 in honour of a
relative who was keen on gambling (Metropolis and
Ulam, 1949).
For notational simplicity, we consider the following
simple model: y = f(x), where the function f represents
the model under study, x:[x1;… xd] is the vector of the
model inputs, and y:[y1;… yn] is the vector of the model
predictions.
Within the context of uncertainty analysis, which is
concerned with estimating the error propagation from a
set of inputs to a set of model outputs, the integral of
interest is the calculation of the mean and variance of the
model outputs which are themselves indeed
multidimensional integrals (the dimensionality number
is determined by the length of the vector of input
parameters):
Eq. 5.38
The goal of an uncertainty analysis is to determine
the uncertainty in the elements of y that results from
uncertainty in the elements of x. Given uncertainty in the
vector x characterised by the distribution functions
D=[D1,… Dd], where D1 is the distribution function
associated with x1, the uncertainty in y is given by:
var ( y ) = ( ( y ) − f ( x ) ) dx
2
E ( y ) = f ( x ) dx
Eq. 5.39
210
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Where, var(y) and E(y) are the variance and expected
value respectively of a vector of random variables, y,
which are computed by the Monte Carlo sampling
technique. In addition to the variance and mean values,
one can also easily compute a percentile for y including
the 95% upper and lower bounds.
5.2.4 Local sensitivity analysis and
identifiability analysis
Forward perturbation:
5.2.4.1 Local sensitivity analysis
Most of the sensitivity analysis results reported in the
literature are of a local nature, and these are also called
one factor at a time (OAT) methods. In OAT methods,
each input variable is varied (also called perturbation)
one at a time around its nominal value, and the resulting
effect on the output is measured. The sensitivity analysis
results from these methods are useful and valid in close
proximity to the parameters analysed, hence the name
local. In addition, the parameter sensitivity functions
depend on the nominal values used in the analysis.
Alternative methods, such as regional or global methods,
expand the analysis from one point in the parameter
space to cover a broader range in the entire parameter
space but this is beyond the scope of this chapter
(interested readers can consult literature elsewhere such
as Saltelli et al., 2000; Sin et al., 2009).
The local sensitivity measure is commonly defined
using the first order derivative of an output, y = f(x), with
respect to an input parameter, x:
Absolute sensitivity: sa =
∂y
∂x
Eq.5.40
(effect on y by perturbing x around its nominal value x0).
Relative sensitivity: sr =
∂y x°
∂x y°
symbolic manipulation toolbox software. Alternatively,
the derivatives can be obtained numerically by model
simulations with a small positive or negative
perturbation, Δx, of the model inputs around their
nominal values, x0. Depending on the direction of the
perturbation, the sensitivity analysis can be
approximated using the forward, backward or central
difference methods:
Eq. 5. 41
(relative effect of y by perturbing x with a fixed fraction
of its nominal value x0).
The relative sensitivity functions are nondimensional with respect to units and are used to
compare the effects of model inputs among each other.
These first-order derivatives can be computed
analytically, for example using Maple or Matlab
0
0
∂y f ( x +Δx ) − f ( x )
=
∂x
Δx
Eq. 5.42
Backward perturbation:
0
0
∂y f ( x ) − f ( x − Δx )
=
∂x
Δx
Eq. 5.43
Central difference:
0
0
∂y f ( x +Δx ) − f ( x − Δx )
=
∂x
2Δx
Eq. 5.44
When an appropriately small perturbation step, Δx, is
selected (usually a perturbation factor, ε = 10-3 is used.
Hence Δx = ε · x), all three methods provide exactly the
same results.
Once the sensitivity functions have been calculated,
they can be used to assess the parameter significance
when determining the model outputs. Typically, large
absolute values indicate high parameter importance,
while a value close to zero implies no effect of the
parameter on the model output (hence the parameter is
not influential). This information is useful to assess
parameter identifiability issues for the design of
experiments.
5.2.4.2 Identifiability analysis using the
collinearity index
The first step in parameter estimation is determining
which sets of parameters can be selected for estimation.
This problem is the subject of identifiability analysis,
which is concerned with identifying which subsets of
parameters can be identified uniquely from a given set of
measurements. Thereby, it is assumed a model can have
a number of parameters. Here the term uniquely is
DATA HANDLING AND PARAMETER ESTIMATION
211
important and needs to be understood as follows: a
parameter estimate is unique when its value can be
estimated independently of other parameter values and
with sufficiently high accuracy (i.e. a small uncertainty).
This means that the correlation coefficient between any
pair of parameters should be low (e.g. lower than 0.5)
and the standard error of parameter estimates should be
low (e.g. the relative error of the parameter estimate, σθ
/θ, lower than e.g. 25%). As it turns out, many parameter
estimation problems are ill-conditioned problems. A
problem is defined ill-conditioned when the condition
number of a function/matrix is very high, which is caused
by multicollinearity issues. In regression problems, the
condition number is used as a diagnostic tool to identify
parameter identifiability issues. Such regression
diagnostics are helpful in generating potential candidates
of the parameter subsets for estimation which the user
can select from.
In Step 1, parameters that have negligible or nearzero influence on the measured model outputs are
screened out from consideration for parameter
estimation. In the second step, for each parameter subset
(all the combinations of the parameter subsets which
include 2, 3, 4,…m parameters) the collinearity index is
calculated. The collinearity index is the measure of the
similarity between any two vectors of the sensitivity
functions. Subsets that have highly similar sensitivity
functions will tend to have a very large number (γK ~ inf),
while independent vectors will have a smaller value γK ~
1 which is desirable. In identifiability analysis, a
threshold value of 5-20 is usually used in literature (Brun
et al., 2001; Sin and Vanrolleghem, 2007; Sin et al.,
2010). It is noted that this γK value is to be used as
guidance for selecting parameter subsets as candidates
for parameter estimation. The best practice is to iterate
and try a number of higher ranking subsets.
There are several identifiability tests suggested in
literature that are entirely based on the sensitivity
functions of the parameters on the outputs. Here we are
using the two-step procedure of Brun et al., 2002.
Accordingly, the procedure works as follows: (i)
assessment of the parameter significance ranking, (ii)
collinearity analysis (dependency analysis of the
parameter sensitivity functions in a parameter subset):
5.3 METHODOLOGY AND WORKFLOW
msqr
Step 1. Rank the significance of the parameters: δ
msqr
δ
=
1
∑N(sri )
N i
Eq. 5.45
Where, sr is a vector of non-dimensional sensitivity
values, sr = i...N values.
Step 2. Calculate the collinearity index of a parameter
subset K, γK.
γ
=
1
minλ
λK = eigen(snormTK snormK )
snorm =
sr
‖sr‖
Eq. 5.46
Eq. 5.47
Eq. 5.48
Where, K indicates a parameter subset, snorm is the
normalized non-dimensional sensitivity function using
the Euclidian norm, and λ represents the eigenvalues of
the normalized sensitivity matrix for parameter subset K.
5.3.1 Data consistency check using an
elemental balance and a degree of
reduction analysis
The following workflow is involved in performing an
elemental balance and a degree of reduction analysis:
Step 1. Formulate a black box process stoichiometry for
the biological process.
In this step, the most relevant reactants and products
consumed and produced in the biological process are
identified and written down. The output is a list of
reactants and products for Step 2.
Step 2. Compose the elemental composition matrices (Em
and Eu).
First establish which variables of interest are measured
and then define the matrices as follows: Em includes the
elemental composition and the degree of reductions for
these measured variables, while Eu includes those of
unmeasured variables. To calculate the degree of
reduction, use the procedure given in Section 5.2.1.2.
Step 3. Compute the unmeasured rates of the species (qu).
Using Em and Eu together with the vector of the measured
rates (qm), the unmeasured rates (qu) are estimated from
the solution of the linear set of equations in Eq. 5.18.
212
Step 4. Calculate the yield coefficients.
In this step, since all the species rates of
consumption/productions are now known, the yield
coefficients can be calculated using Eq. 5.2 and the
process stoichiometry can be written using the yield
coefficient values.
Step 5. Verify the elemental balance.
In this step, a simple check is performed to verify if the
elemental balance and degree of reduction balance are
closed. If not, the procedure needs to be iterated by
assuming a different hypothesis concerning the
formation of by-products.
5.3.2 Parameter estimation workflow for
the non-linear least squares method
This workflow assumes that an appropriate and
consistent mathematical model is used to describe the
data. Such a model confirms the elemental balance and
degree of reduction analysis (see the workflow in Section
5.3.1). Usually these models are available from
literature. Most of them are modified from ASM models
with appropriate simplifications and/or additions
reflecting the conditions of the batch experiment.
Step 1. Initialisation.
In this step, the initial conditions for the model variables
are specified as well as a nominal set of parameters for
the model. The initial conditions for the model are
specified according to the experimental conditions (e.g.
10 mg NH4-N added at time 0, kLa is a certain value,
oxygen saturation at a given temperature is specified,
etc.). An initial guess of the model parameters is taken
from literature.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
In this step, the parameter estimation problem is defined
as a minimization problem and solved using optimization
algorithms (e.g. fminsearch in Matlab)
Step 4. Estimate the uncertainty of the parameter
estimators and model outputs.
In this step, calculate the covariance matrix of the
parameter estimators and compute the parameter
confidence intervals as well as the parameter correlation
matrix. Given the covariance matrix of the parameter
estimators, estimate the covariance matrix of the model
outputs by linear error propagation.
Step 5. Review and analyse the results.
In this step, review the values of the parameter values,
which should be within the range of parameter values
obtained from the literature. In addition, inspect the
confidence intervals of the parameter estimators. Very
large confidence intervals imply that the parameter in
question may not be estimated reliably and should be
excluded from the subset.
Further, plot and review the results from the best-fit
solution. Typically, the data and model predictions
should fit well.
If the results (both parameter values) and the best fit
solution to the data are not satisfactory, iterate as
appropriate by going back to Step 1 or Step 2.
5.3.3 Parameter estimation workflow for
the bootstrap method
The workflow of the bootstrap method follows on from
Step 1, Step 2 and Step 3 of the non-linear least squares
method.
Step 2. Select the experimental data and a parameter
subset for the parameter estimation.
Step 1. Perform a reference parameter estimation using
the non-linear least squares method.
In this step, the experimental data is reviewed for the
parameter estimation and which parameters need to be
estimated is defined. This can be done using expert
judgement or, more systematically, a sensitivity and
identifiability analysis (see Section 5.3.4).
This step is basically an execution of steps 1, 2 and 3 of
the workflow in the non-linear least squares technique.
The output is a residual vector that is passed on to the
next step. The residual vector is then plotted and
reviewed. If the residuals follow a systematic pattern (it
should be random) or contain outliers, this is a cause for
concern as it may imply the bootstrap method is not
suited for this application.
Step 3. Define and solve the parameter estimation
problem.
DATA HANDLING AND PARAMETER ESTIMATION
213
Step 2. Generate synthetic data by bootstrap sampling and
repeat the parameter estimation.
forward, backward or central difference. Plot, review and
analyse the results.
Synthetic data is generated using Eq. 5.29-5.32 by
performing bootstrap sampling (random sampling with
replacement) from the residual vector and adding it to the
model prediction obtained in Step 1. For each synthetic
data, the parameter estimation in Step 1 is repeated and
the output (that is, the values of the parameter estimators)
is recorded in a matrix.
Step 3. Rank the parameter significance.
Calculate the delta mean-square measure, δmsqr, and rank
the parameters according to this measure. Exclude any
parameters that have zero or negligible impact on the
outputs.
Step 4. Compute the collinearity index.
Step 3. Review and analyse the results.
In this step, the mean, standard deviation and the
correlation matrix of the parameter estimators are
computed from the recorded matrix data in Step 2.
Moreover, the distribution function of the parameter
estimators can be estimated and plotted using the vector
of the parameter values that was obtained in Step 2.
For all the parameter combinations (e.g. subset size 2, 3,
4….m), the collinearity index, γK, is calculated. Each
parameter subset is ranked according to the collinearity
index value.
Step 5. Review and analyse the results.
As in Step 5 of the workflow in the non-linear least
squares method, the results are interpreted and evaluated
using knowledge from literature and process
engineering.
Based on the results from Step 3 and Step 4, identify a
short list of candidates (parameter subsets) that are
identifiable. Exclude these parameters from any
parameter subset that has near-zero or negligible
sensitivity on the outputs.
5.3.4 Local sensitivity and identifiability
analysis workflow
5.3.5 Uncertainty analysis using the Monte
Carlo method and linear error propagation
The workflow of this procedure starts with the
assumption that a mathematical model is available and
ready to be used to describe a set of experimental data.
The workflow for the Monte Carlo method includes the
following steps:
Step 1. Input the uncertainty definition.
Step 1. Initialisation.
A framework is defined for the sensitivity analysis by
defining the experimental conditions (the initial
conditions for the batch experiments) as well as a set of
nominal values for the model analysis. The model is
solved with these initial conditions and the model outputs
are plotted and reviewed before performing the
sensitivity analysis.
Step 2. Compute the sensitivity functions.
Define which outputs are measured and hence should be
included in the sensitivity analysis. Define the
experimental data points (every 1 min versus every 5
min).
Compute the sensitivity functions of the parameters
on the outputs using a numerical difference, e.g. using a
Identify which inputs (parameters) have uncertainty.
Define a range/distribution for each uncertainty input,
e.g. normal distribution, uniform distribution, etc. The
output from the parameter estimators (e.g. bootstrap) can
be used as input here.
Step 2. Sampling from the input space.
Define the sampling number, N, (e.g. 50, 100, etc.) and
sample from the input space using an appropriate
sampling technique. The most common sampling
techniques are random sampling, Latin Hypercube
sampling, etc. The output from this step is a sampling
matrix, XNxm, where N is the number of samples and m is
the number of inputs.
214
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Step 3. Perform the Monte Carlo simulations.
Perform N simulations with the model using the
sampling matrix from Step 2. Record the outputs in an
appropriate matrix form to be processed in the next step.
Step 4. Review and analyse the results.
Plot the outputs and review the results. Calculate the
mean, standard deviation/variance, and percentiles (e.g.
95 %) for the outputs. Analyse the results within the
context of parameter estimation quality and model
prediction uncertainty. Iterate the analysis, if necessary,
by going back to Step 1 or Step 2.
plus a degree of reduction balance), measurement of
three rates is sufficient to estimate/infer the remaining
rates.
To illustrate the concept, the measured rates are
selected as the volumetric consumption rate of substrate
(-qs), the biomass production rate (qx), and the glycerol
production rate (qg) hence the remaining rates for
ammonia consumption as well as the production of
ethanol and CO2 need to be estimated using Eq. 5.18. In
the measured rate vectors, a negative sign indicates the
consumption of a species, while a positive sign indicates
the production of a species.
Step 3 Compute the unmeasured rates of the species (qu).
The workflow for linear error propagation:
Recall Eq. 5.18, which is solved as follows:
The workflow is relatively straightforward as it is
complementary to the covariance matrix of the parameter
estimators and should be performed as part of the
parameter estimation in the non-linear least squares
method. It requires the covariance matrix of parameter
estimators as well as the Jacobian matrix which are both
obtained in Step 4 of the non-linear least squares
methodology.
5.4 ADDITIONAL EXAMPLES
E m ⋅ q m + Eu ⋅ qu = 0
S
X
Gly
C 1
1
1
N 0 0.15
0
γ 4 4.12 4.67
NH 3
Eth CO 2
-q s 0 1.0 1.0 -q n
⋅ q + 1 0
0 ⋅ q e = 0
x
q g 0 6
0 q c
q u = − ( Eu ) ⋅ E m ⋅ q m
−1
Example 5.2 Anaerobic fermentation of glucose
−1
In this example, anaerobic fermentation of glucose to
ethanol and glycerol as metabolic products is considered.
0 1.0 1.0 1
1
1
-q n
⋅ 0 0.15
q
=
−
1
0
0
0
e
q
0 6
0 4 4.12 4.67
c
-q s
⋅ q
x
q g
Step 1. Formulate the process stoichiometry.
Ammonia is assumed to be the nitrogen source for
growth. The biomass composition is assumed to be
CH1.6O0.5N0.15. All the substrates are given on the basis
of 1 C-mol, whereas nitrogen is on the basis of 1 N-mol.
In this biological process, the substrates are CH2O
(glucose) and NH3. The products are CH1.61O0.52N0.15
(biomass), CH3O0.5 (ethanol), CH8/3O (glycerol) and CO2.
Water is excluded from the analysis, as its rate of
production is not considered relevant to the process. This
means that the H and O balances will not be considered
either.
Solving the system of linear equations above yields
the following solution where the three unmeasured rates
are calculated as a function of the measured rates qs, qg
and qx:
Step 2. Compose the elemental composition matrices (Em
and Eu).
Once the rates of all the products and substrates are
estimated, one can then calculate the yield coefficients
for the process by recalling Eq. 5.2 as follows:
As the process has six species (substrates + products) and
three constraints (two elemental balances for C and N
-q n − 0.15q x
q e = 2q s / 3 − 467q g / 600 − 103q x / 150
q q / 3 − 133q / 600 − 47q / 150
c s
g
x
Step 4. Calculate the process yields.
DATA HANDLING AND PARAMETER ESTIMATION
q
Ysx = x
qs
and Ysg =
215
qg
qs
−0.15q x
q
Ysn = n =
= − 0.15Ysx
qs
qs
Yse =
qe ( 2qs / 3 − 467qg / 600 − 103qx / 150)
=
=
qs
qs
2
− 0.7783Ysg − 0.6867Ysx
3
q ( qs / 3 − 133qg / 600 − 47qx / 150)
=
Ysc = c =
qs
qs
1
− 0.2217Ysg − 0.3133Ysx
3
With these yield coefficients estimated,
simplified process stoichiometry reads as follows:
the
0 = − CH2O − 0.15Ysx ⋅ NH3…
2
+ Ysx ⋅ CH1.61O0.52 N0.15 + − 0.7783Ysg − 0.6867Ysx ⋅ CH3O0.5 +
3
1
Ysg ⋅ CH8/3O + − 0.2217Ysg − 0.3133Ysx ⋅ CO2
3
Step 5. Verify the elemental balance.
From the process stoichiometry, it is straightforward to
verify that the elemental and degree of reduction
balances are closed:
2
−1 + Ysx + − 0.7783 ⋅ Ysg − 0.6867 ⋅ Ysx + Ysg +
3
1
− 0.2217 ⋅ Ysg − 0.3133 ⋅ Ysx = 0
3
The nitrogen balance:
−0 − 0.15 ⋅ Ysx + 0.15 ⋅ Ysx + 0 + 0 + 0 = 0
The degree of reduction balance:
2
−1 ⋅ 4 + Ysx ⋅ 4.12 + − 0.7783 ⋅ Ysg − 0.6867 ⋅ Ysx ⋅ 6 +
3
Ysg ⋅ 4.67 = 0
In the above example, three measured rates were
assumed available as a minimum requirement to identify
the system of linear equations. In practical applications,
there might be two other situations: (i) redundant
measurements: measurements of most or perhaps all of
the species rates of production/consumption are
available. In this case, the additional rate measurements
can be used for data quality check and validation (where
some of the measured rates could be used as a validation
of the estimated coefficients from the balance analysis);
(ii) a limited set of measurements: in this case too few
rates measurements are available to uniquely estimate
the unmeasured rates. If data from enough variables is
not available, some variables should be fixed (e.g. this
could be done iteratively in a search algorithm) to reduce
the degrees of freedom. If not, there are infinite solutions.
Further discussion of these techniques can be found
elsewhere in relevant literature (Villadsen et al., 2011;
Meijer et al., 2002; van der Heijden et al., 1994).
Example 5.3 Estimate the parameters of the ammonium and
nitrite oxidation processes using data from batch tests: the nonlinear least squares method
Aerobic batch tests with a sludge sample from a predenitrification plant are performed to measure the
parameters of the nitrifying bacteria, in particular the
ammonium-oxidizing organisms (AOO) and nitriteoxidising
organisms
(NOO).
Following
the
recommended experimental procedure in literature
(Guisasola et al., 2005), two separate batch tests were
performed as follows: (i) batch test 1 with added
ammonium of 20 mg NH4-N L-1 and an inhibitor (sodium
azide) to suppress NOO activity, (ii) batch test 2 with
added ammonium of 20 mg NH4-N L-1 without any
inhibitor addition. In both tests, both the pH and
temperature are controlled at 7.5 and 25 ºC respectively.
During both the batch tests, ammonium, nitrite and
nitrate are measured every 5 minutes, while dissolved
oxygen is measured every minute. The data collected is
shown in Figure 5.2. For the sake of simplicity and to
keep the focus on the demonstration of the methods and
their proper interpretation, these examples use synthetic
data with random (white-noise) addition.
Part 1. Estimate the parameters of the ammonium oxidation
process.
Several models are suggested in literature reviewed in
Sin et al., 2008. We use the following mathematical
model given in Table 5.1 to describe the kinetics of
ammonium and nitrite oxidation. For the sake of
simplicity, the following is assumed: (i) endogenous
216
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
respiration related to heterotrophic biomass is constant
(hence not modelled), (ii) the inert fraction of the
biomass released during decay is negligible (hence not
modelled), and (iii) the ammonium consumed for
autotrophic growth of biomass is negligible. It is noted
that for the sake of completeness all of the above
phenomena should be described which makes the
analysis more accurate. However, here the model is kept
simple to focus the attention of the reader on the
workflow of parameter estimation.
Table 5.1 The two-step nitrification model structure using matrix representation (adopted from Sin et al., 2008)
Variables→Ci
Processes ↓j
SNH
mg N L-1
SO
mg O2 L-1
‒1
YAOO
AOO growth
1‒
3.43
YAOO
SNO3
mg N L-1
1
YAOO
1
AOO decay
1‒
NOO growth
1.14
YNOO
‒1
YNOO
XAOO
mg COD L-1
XNOO
mg COD L-1
Rates
qj
1
µmaxAOO · MNH · MO,AOO · XAOO
‒1
bAOO · XAOO
1
YNOO
1
1
NOO decay
Aeration
MNH :
SNO2
mg N L-1
1
µmaxNOO · MNO2 · MO,NOO · XNOO
‒1
bNOO · XNOO
kLa·(Sosat ‒ So)
SNH
SO
SO
SNO2
:
:
:
;M
;M
;M
SNH +Ks,AOO O,AOO SO +Ko,AOO O,NOO SO +Ko,NOO NO2 SNO2 +Ks,NOO
The model has in total six ordinary differential
equations (ODE), which corresponds to one mass
balance for each variable of interest. Using a matrix
notation, each ODE can be formulated as follows:
dCi
= ν ij ⋅ q j
dt
j
Eq. 5.49
The model is implemented in Matlab and solved
using a standard differential equation solver (ODE45 in
Matlab).
%% solve the ODE model:
%[time,output] = ODEsolver('Model',[starttime
simulation
The model has six state variables, all of which need
to be specified to solve the system of the ODE equations.
The initial condition corresponding to batch test 1 is
shown in Table 5.3.
Step 2. Select the measurements and parameter subset for
parameter estimation.
We used data collected from batch test 1 which includes
ammonium, nitrite and dissolved oxygen measurements.
Due to the suppression of NOO activity, no nitrate
production is observed. Since batch test 1 is not designed
for decay rate coefficient estimation, we consider all the
parameters of AOO except bAOO for estimation. Hence,
the following is our selection:
endtime simulation],Initial conditions for
model variables,simulation options,model parameters);
options=odeset('RelTol',1e-7,'AbsTol',1e-8);
[t,y] = ode45(@nitmod,t,x0,options,par);
•
•
Y = [NH4 NO2 DO]; selected measurement set, Y.
θ = [YAOO µmaxAOO Ks,AOO Ko,AOO]; parameter subset
for the estimation.
Step 1. Initialisation.
Step 3. Solve the parameter estimation problem.
The model has in total 12 parameters. The nominal
values as well as their range are taken from literature (Sin
et al., 2008) and shown in Table 5.2
The parameter estimation is programmed as a
minimization problem using the sum of the squared
errors as the cost function and solved using an
unconstrained
non-linear
optimisation
solver
DATA HANDLING AND PARAMETER ESTIMATION
217
(fminsearch algorithm in Matlab) using the initial
parameter guess given in Table 5.2 and initial conditions
in Table 5.3. To simulate the inhibitor addition, the
maximum growth rate of NOO is assumed to be zero in
the model simulations. The best estimates of the
parameter estimators are given in Table 5.3.
Table 5.2 Nominal values of the model parameters used as an initial guess for parameter estimation together with their upper and lower bounds.
Parameter
Ammonium-oxidising organisms (AOO)
Biomass yield
Maximum growth rate
Substrate (NH4) affinity
Oxygen affinity
Decay rate coefficient
Nitrite-oxidising organisms (NOO)
Biomass yield
Maximum growth rate
Substrate (NO2) affinity
Oxygen affinity
Decay rate coefficient
Experimental setup
Oxygen mass transfer
Oxygen saturation
Symbol
Unit
Nominal value
Range
YAOO
mg COD mg N-1
d-1
mg N L-1
mg O2 L-1
d-1
0.15
0.8
0.4
0.5
0.1
0.11 - 0.21
0.50 - 2.10
0.14 - 1.00
0.10 - 1.45
0.07 - 0.30
mg COD mg N-1
d-1
mg N-1
mg O2 L-1
d-1
0.05
0.5
1.5
1.45
0.12
0.03 - 0.09
0.40 - 1.05
0.10 - 3.00
0.30 - 1.50
0.08 - 0.20
d-1
mg O2 L-1
360
8
*
*
AOO
max
μ
Ks,AOO
Ko,AOO
bAOO
YNOO
NOO
max
μ
Ks,NOO
Ko,NOO
bNOO
kL a
SO,sat
* Not estimated in this example but assumed known.
Table 5.3 Initial condition of the state variables for the model in batch test 1.
Variable
Ammonium
Oxygen
Nitrite
Nitrate
AOO biomass
NOO biomass
Symbol
SNH
SO
SNO2
SNO3
XAOO
XNOO
Unit
mg N L-1
mg O2 L-1
mg N L-1
mg N L-1
mg COD L-1
m COD L-1
Initial value
20
8
0
0
75
25
9
25
20
NH4+-N
8
NO2--N
7
DO (mg O2 L-1)
NH4+-N, NO3--N, NO2--N (mg N L-1)
Comment
Pulse addition
Saturation
Post denitrified
Post denitrified
Ratio of AOO to NOO reflects ratio of their yields
NO3--N
15
10
6
5
4
3
2
5
1
0
0.0
1.2
2.4
3.6
0
0.0
Time (h)
Figure 5.2 Data collected in batch test 1. NH4, NO2 and DO are used as the measured data set.
1.2
2.4
Time (h)
3.6
218
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
%%step 3 define and solve parameter estimation
%% get the Jacobian matrix. use built-in
problem (as a minimization problem)
"lsqnonlin.m" but with no iteration.
options =optimset('display',
options =optimset('display',
'iter','tolfun',1.0e-06, 'tolx',1.0e-5,
'iter','tolfun',1.0e-06, 'tolx',1.0e-5,
'maxfunevals', 1000);
'maxfunevals', 0);
[pmin,sse]=fminsearch(@costf,pinit,options,td,yd
,idx,iy);
[~,~,residual,~,~,~,jacobian]=lsqnonlin(@costl,p
Step 4. Estimate the uncertainty of the parameter
estimators and the model prediction uncertainty.
In this step, the covariance matrix of the parameter
estimators is computed. From the covariance matrix, the
standard deviation, 95 % confidence interval as well as
the correlation matrix are obtained. The results are
shown in Table 5.4.
min,[],[],options,td,yd,idx,iy);
j(:,:)=jacobian; e=residual;
s=e'*e/dof; %variance of errors
%% calculate the covariance of parameter
estimators
pcov = s*inv(j'*j) ; %covariance of parameters
psigma=sqrt(diag(pcov))'; % standard deviation
parameters
pcor = pcov ./ [psigma'*psigma]; % correlation
matrix
alfa=0.025; % significance level
Table 5.4 Optimal values of the parameter estimators after the solution of
the parameter estimation problem.
tcr=tinv((1-alfa),dof); % critical t-dist value
at alfa
p95 =[pmin-psigma*tcr; pmin+psigma*tcr]; %+-95%
Initial guess, θ°
0.1
0.8
0.4
0.5
Parameter
YAOO
μmaxAOO
Ks,AOO
Ko,AOO
Optimal values, θ
0.15
1.45
0.50
0.69
confidence intervals
Table 5.5 Parameter estimation quality for the ammonium oxidation process: standard deviation, 95% confidence intervals and correlation matrix.
Parameter
YAOO
μmaxAOO
Ks,AOO
Ko,AOO
Optimal value, θ
0.15
1.45
0.50
0.69
Standard deviation,
0.0076
0.0810
0.0180
0.0590
95 % confidence interval (CI)
0.130
1.290
0.470
0.570
Using the covariance matrix of the parameter
estimators, the uncertainty in the model prediction is also
calculated and the results are shown in Figure 5.3.
0.160
1.610
0.540
0.800
YAOO
1
Correlation matrix
μmaxAOO
Ks,AOO
0.96
0.0520
1
0.0083
1
Ko,AOO
0.17
0.42
-0.26
1
%% calculate confidence intervals on the model
output
ycov = j * pcov * j';
ysigma=sqrt(diag(ycov)); % std of model outputs
ys=reshape(ysigma,n,m);
y95 = [y(:,iy) - ys*tcr y(:,iy)+ys*tcr]; % 95%
confidence intervals
219
20
20
18
18
16
16
14
14
8
7
12
10
8
12
10
8
6
6
4
4
2
2
0
DO (mg O2 L-1)
-1
NO2--N (mg N L )
NH4+-N (mg N L-1)
DATA HANDLING AND PARAMETER ESTIMATION
1.2
2.4
Time (h)
3.6
5
4
3
0
0
6
2
0
1.2
2.4
Time (h)
3.6
1.2
2.4
Time (h)
3.6
Figure 5.3 Model outputs including 95 % confidence intervals calculated using linear error propagation (red lines). The results are compared with the
experimental data set.
Step 5. Review and analyse the results.
The estimated parameter values (Table 5.5) are found to
be within the range reported in literature. This is an
indication that the parameter values are credible. The
uncertainty of these parameter estimators is found to be
quite low. For example, the relative error (e.g. standard
deviation/mean value of parameter values) is less than 10
%, which is also reflected in the small confidence
interval. This indicates that the parameter estimation
quality is good. It is usually noted that relative error
higher than 50 % is indicative of bad estimation quality,
while relative error below 10 % is good.
Regarding the correlation matrix, typically from
estimating parameters from batch data for Monod-like
models, the growth yield is significantly correlated with
the maximum growth rate (the linear correlation
coefficient is 0.96). Also notable is the correlation
between the maximum growth rate and the oxygen
affinity constant. This means that a unique estimation of
the yield and maximum growth rate is not possible.
Further investigation of the correlation requires a
sensitivity analysis, which is demonstrated in Example
5.5.
Since the parameter estimation uncertainty is low, the
uncertainty in the model predictions is also observed to
be small. In Figure 5.3, the mean (or average) model
prediction and the 95 % upper and lower bounds are quite
close to each other. This means that the model prediction
uncertainty due to parameter estimation uncertainty is
negligible. It is noted that a comprehensive uncertainty
analysis of the model predictions will require analysis of
all the other sources of uncertainty including other model
parameters as well as the initial conditions. However,
this is outside the scope of this example and can be seen
elsewhere (Sin et al., 2010). Measurement error
uncertainty is considered in Example 5.6.
This concludes the analysis of parameter estimation
using the non-linear least squares method for the AOO
parameters.
Part 2. Estimate the parameters for the NOO step.
Steps 1 and 2. Initial conditions and selection of data and
parameter subsets for the parameter estimation.
The same initial condition for batch test 1 is used in batch
test 2 but without any inhibitor addition, meaning that in
this example the nitration is active. The data collected
from batch test 2 is shown in Figure 5.3, which includes
ammonium, nitrite, nitrate and DO measurements.
•
Y2 = [NH4 NO2 NO3 DO]; selected measurement set,
Y.
220
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
•
The parameter values of AOO were set to the
estimated values (Table 5.4) in the first part and are
hence known, while the yield and kinetic parameters of
NOO can be identified from the data:
8
20
7
NH4+-N
O2
NO2--N
NO3--N
15
6
DO (mg O2 L-1)
NH4+-N, NO3--N, NO2--N (mg N L-1)
θ2 = [YNOO µmaxNOO Ks,NOO Ko,NOO]; parameter subset
for the estimation.
10
5
4
5
3
0
2
0
1.2
2.4
Time (h)
3.6
0
4.8
1.2
2.4
Time (h)
3.6
4.8
Figure 5.4 Measured data from batch test 2.
Steps 3 and 4. Solve the parameter estimation problem and
calculate the parameter estimation uncertainties.
problem as well as the parameter uncertainties for NOO
are shown in Table 5.6.
The AOO parameters are previously estimated in Part 1.
The results of the solution of the parameter estimation
.
Table 5.6 Optimal values of the parameter estimators after solution of the parameter estimation problem.
Parameter
YNOO
μmaxNOO
Ks,NOO
Ko,NOO
Optimal values, θ
0.04
0.41
1.48
1.50
Standard deviation,
0.01
0.13
0.03
0.05
95 % confidence interval (CI)
Lower bound Upper bound
0.01
0.07
0.15
0.66
1.42
1.55
1.39
1.60
The linear propagation of the parameter estimation
error (covariance matrix) to the model prediction
uncertainty is shown in Figure 5.5.
YNOO
1.00
Correlation matrix
μmaxNOO
Ks,NOO
1.00
0.54
1.00
0.55
1.00
Ko,NOO
-0.86
-0.86
-0.37
1.00
DATA HANDLING AND PARAMETER ESTIMATION
221
20
Nitrate (mg N L-1)
-1
Ammonium (mg N L )
20
10
10
NO2--N
NH4+-N
0
0.0
1.2
2.4
Time (h)
3.6
0
0.0
4.8
20
1.2
2.4
Time (h)
3.6
8
DO
DO (mg O2 L )
-1
-1
Nitrate (mg N L )
NO3--N
10
0
0.0
4.8
6
4
2
1.2
2.4
Time (h)
3.6
4.8
0.0
1.2
2.4
Time (h)
3.6
4.8
Figure 5.5 Model outputs including 95 % confidence intervals compared with the experimental data set.
Step 5. Review and analyse the results
The estimated parameter values are within the range
reported for the NOO parameters in literature, which
makes them credible. However, this time the parameter
estimation error is noticeably higher, e.g. the relative
error (the ratio of standard deviation to the optimal
parameter value) is more than 30%, especially for the
yield and maximum growth rate. This is not surprising
since the estimation of both the yield and maximum
growth rate is fully correlated (the pairwise linear
correlation coefficient is 1). These statistics mean that a
unique parameter estimation for the yield, maximum
growth rate and oxygen half-saturation coefficient of
NOO (the pairwise linear correlation coefficient is 0.86)
is not possible with this batch experiment. Hence, this
parameter subset should be considered as a subset that
provides a good fit to the experimental data, while
individually each parameter value may not have
sensible/physical meaning.
The propagation of the parameter covariance matrix
to the model prediction uncertainty indicates low
uncertainty on the model outputs. This means that
although parameters themselves are not uniquely
identifiable, they can still be used to perform model
predictions, e.g. to describe batch test data. While
performing simulations with the model, however, one
needs to report the 95% confidence intervals of the
simulated values as well. The latter reflects how the
covariance of the parameter estimates (implying the
parameter estimation quality) affects the model
prediction quality. For example, if the 95% confidence
interval of the model predictions is low, then the effect
of the parameter estimation error is negligible.
Part 1 and Part 2 conclude the parameter estimation
for the two-step nitrification step. The results show that
the quality of the parameter estimation for AOO is
relatively higher than that of NOO using batch data for
these experiments. This poor identifiability will be
investigated later on, using sensitivity analysis to
improve the identifiability of individual parameters of
the model.
Regarding the model prediction errors, the 95%
confidence interval of the model outputs is quite low.
This means that the effects of the parameter estimation
errors on the model outputs are low.
222
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Example 5.4 Estimate the parameters of ammonium oxidation
using data from the batch test – the bootstrap method
Step 1. Perform a reference parameter estimation using
non-linear least squares.
In this example, we investigate the parameter estimation
problem in part 1 of Example 5.3. We used the data from
batch test 1 to estimate the parameters of AOO.
The workflow in this step is exactly the same as the steps
1, 2 and 3 in Example 5.3. The output from this step is
the best fit to the data and the distribution of residuals
(Figure 5.6).
0.2
Residuals
Ammonium
0.0
-0.2
0
Residuals
0.2
5
10
15
20
25
30
35
5
10
15
20
25
30
35
5
10
15
20
Data index
25
30
35
Nitrite
0.0
-0.2
0
Residuals
0.2
Oxygen
0.0
-0.2
0
Figure 5.6 Residuals from the reference parameter estimation.
Step 2. Generate synthetic data by bootstrap sampling and
repeat the parameter estimation.
ybt = y(:,iy) + rsam ; % synthetic data: error
+ model (ref PE)
options
In this step, bootstrap sampling from residuals is
performed.
=optimset('display','iter','tolfun',1.0e06,'tolx',1.0e-5,'maxfunevals',1000);
[pmin(i,:),sse(i,:)]=lsqnonlin(@costl,pmin1,plo,
nboot=50; % bootstrap samples
phi,options,td,ybt,idx,iy);
for i=1:nboot
bootsam(:,:,i)=ybt; % record samples
end
disp(['the iteration number is :
',num2str(i)])
onesam =ceil(n*rand(n,m)); % random sampling
with replacement
rsam
=res(onesam); % measurement errors for
each variable
Fifty bootstrap samples from residuals (random
sampling with replacement) are performed and added to
the model, thereby yielding the 50 synthetic
measurement data sets shown in Figure 5.7.
Nitrate (mg N L-1)
Ammonium (mg N L-1)
DATA HANDLING AND PARAMETER ESTIMATION
20
10
0
0.00
0.02
0.04
0.06
0.08
0.10
1.20
0.02
0.04
0.06
0.08
0.10
1.20
0.02
0.04
0.06
0.08
0.10
1.20
20
10
0
0.00
Oxygen (mg O2 L-1)
223
10
5
0
0.00
Time (h)
Figure 5.7 Generation of synthetic data using bootstrap sampling from the residuals (50 samples in total).
For each of this synthetic data (a bootstrap sample),
a parameter estimation is performed and the results are
recorded for analysis. Because 50 synthetic data sets are
generated, this means that 50 different estimates of
parameters are obtained. The results are shown as a
histogram for each parameter estimate in Figure 5.8.
20
30
Counts
Counts
20
10
10
0
0.12
0.13
0.14
0.15
0.16
0.17
0
1.0
0.18
1.1
1.2
20
1.4
1.5
1.6
1.7
1.8
0.80
0.85 0.90
15
Counts
Counts
1.3
Distribution of µAOO
max
Distribution of YAOO
10
10
5
0
0.45
0.47
0.5
0.52
Distribution of KS,AOO
0.55
2
0.50
0.55 0.60
0.65
0.70
0.75
Distribution of KO,AOO
Figure 5.8 Distribution of the parameter estimates obtained using the bootstrap method (each distribution contains 50 estimated values for each
parameter).
224
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
.%step 3 Evaluate/interpret distribution of theta
Step 3. Review and analyse the results.
disp('The mean of distribution of theta are')
Step 2 provided a matrix of the parameter estimates,
θ50x4. In this step, the mean, standard deviation and
correlation matrix properties of this matrix are evaluated.
The results are shown in Table 5.7.
disp(mean(pmin))
disp('The std.dev. of distribution of theta
are')
disp(std(pmin))
disp('')
disp('The correlation of parameters')
disp(corr(pmin))
Table 5.7 Optimal values of the parameter estimators after solving the parameter estimation problem.
Parameter
YAOO
μmaxAOO
Ks,AOO
Ko,AOO
Optimal value, θ
0.14
1.40
0.50
0.68
Standard deviation,
0.01
0.11
0.02
0.07
All the results, including the mean parameter
estimates, their standard deviation and the correlation
matrix are in good agreement with the parameter
estimates obtained from the non-linear least squares
method (compare with Table 5.6.). This is expected,
since the distribution of residuals is found to be quite
similar to a normal distribution (Figure 5.6). In this case,
both the non-linear least squares (and the linear
approximation of covariance matrix estimation) as well
as the bootstrap method will obtain statistically similar
results.
Also the model simulation with the mean values
obtained from the bootstrap samples provided similarly
good fit to the measured data, as shown in Figure 5.3.
Because the bootstrap method is intuitively simple
and straightforward and does not require a calculation of
the Jacobian matrix, we recommend it for practical use.
However, a reservation on using this method is that the
distribution of the residuals should be inspected and
should not contain any systematic pattern (indicating
model structure or systematic measurement issues).
%% get the Jacobian matrix. use built-in
"lsqnonlin.m" but with no iteration.
options =optimset('display',
'iter','tolfun',1.0e-06, 'tolx',1.0e-5,
'maxfunevals', 0);
YAOO
1.00
Correlation matrix
μmaxAOO
Ks,AOO
0.97
-0.03
1.00
-0.07
1.00
Ko,AOO
0.20
0.41
-0.28
1.00
[~,~,residual,~,~,~,jacobian]=lsqnonlin(@costl,p
min,[],[],options,td,yd,idx,iy);
j(:,:)=jacobian; e=residual;
s=e'*e/dof; %variance of errors
%% calculate the covariance of parameter
estimators
pcov = s*inv(j'*j) ; %covariance of parameters
psigma=sqrt(diag(pcov))'; % standard deviation
parameters
pcor = pcov ./ [psigma'*psigma]; % correlation
matrix
alfa=0.025; % significance level
tcr=tinv((1-alfa),dof); % critical t-dist value
at alfa
p95 =[pmin-psigma*tcr; pmin+psigma*tcr]; %+-95%
confidence intervals
Example 5.5 Sensitivity and identifiability analysis of the
ammonium oxidation process parameters in batch tests
Here the ammonium oxidation process is used as
described in Example 5.3. The objective of this example
is twofold: in the first part, we wish to assess the
sensitivity of all the AOO parameters to all the model
outputs under the experimental conditions of batch test
1. In the second part we wish to examine, given the
measured data set, which parameter subsets are
potentially identifiable and compare them with the
parameter subset already used in the parameter
estimation in Example 5.3.
DATA HANDLING AND PARAMETER ESTIMATION
Step 1. Initialisation. We use the initial conditions of
batch test 1 as described in Table 5.3 as well as the
nominal values of AOO model parameters as given in
Table 5.2.
The model outputs of interest are:
•
y = [NH4 NO2 NO3 DO AOO NOO]
The parameter set of interest is:
•
θ = [YAOO µmaxAOO Ks,AOO Ko,AOO bAOO]
Step 2. Compute and analyse the sensitivity functions.
In this step, the absolute sensitivity functions are
computed using numerical differentiation and the results
are recorded for analysis.
for i=1:m; %for each parameter
dp(i) = pert(i) * abs(ps(i)); % parameter
perturbation
p(i)
= ps(i) + dp(i);
225
the yield and ammonium (substrate) consumption. A
higher yield means less ammonium is consumed per unit
growth of biomass, and hence it would also mean more
ammonium present in the batch test. Since less
ammonium is consumed, less nitrite would be produced
(hence the negative correlation).
On the other hand, it is also noted that the sensitivity
of the yield parameter increases gradually during the
linear growth phase and starts to decrease as we are
nearer to the depletion of ammonium. Once the
ammonium is depleted, the sensitivity becomes nil as
expected. As predicted, the yield has a positive impact
on AOO growth since a higher yield means higher
biomass production. Regarding oxygen, the yield first
has a positive impact that becomes negative towards the
completion of ammonium. This means there is a rather
non-linear relationship between the oxygen profile and
the yield parameter. As expected, the yield of AOO has
no impact on the nitrate and NOO outputs in batch test 1,
because of the addition of the inhibitor that effectively
suppressed the second step of nitrification.
% forward
perturbation
[t1,y1] = ode45(@nitmod,td,x0,options,p);
p(i) = ps(i) - dp(i); %backward perturbation
[t2,y2] = ode45(@nitmod,td,x0,options,p);
dydpc(:,:,i) = (y1-y2) ./ (2 * dp(i));
%central difference
dydpf(:,:,i) = (y1-y) ./ dp(i); %forward
difference
dydpb(:,:,i) = (y-y2) ./ dp(i); %backward
difference
p(i)=ps(i); % reset parameter to its reference
value
end
The output sensitivity functions (absolute) are
plotted in Figure 5.9 for one parameter, namely the yield
of AOO growth for the purpose of detailed examination.
The interpretation of a sensitivity function is as follows:
(i) higher magnitude (positive or negative alike) means
higher influence, while lower or near zero magnitude
means negligible/zero influence of the parameter on the
output, (ii) negative sensitivity means that an increase in
a parameter value would decrease the model output, and
(iii) positive sensitivity means that an increase in a
parameter value would increase the model output.
With this in mind, it is noted that the yield of AOO has a
positive effect on ammonium and an equally negative
impact on nitrite. This is expected from the model
structure where there is an inverse relationship between
In the sensitivity analysis, what is informative is to
compare the sensitivity functions among each other. This
is done in Figure 5.10 using non-dimensional sensitivity
functions, which are obtained by scaling the absolute
sensitivity function with their respective nominal values
of parameters and outputs (Eq. 5.41). Figure 5.10 plots
the sensitivity of all the model parameters with respect to
the six model outputs. Each subplot in the figure presents
the sensitivity functions of all the parameters with
respect to one model output shown in the legend. The yaxis indicates the non-dimensional sensitivity measure,
while the x-axis indicates the time during the batch
activity. For example, we observe that the sensitivity of
parameters to nitrate and NOO is zero. This is logical
since NOO activity is assumed to be zero in this
simulation.
For the model outputs for ammonium, nitrite and
oxygen, the sensitivity functions of the yield and
maximum growth rate for AOO follow an inversely
proportional trend/pattern. This inversely proportional
relation is the reason why the parameter estimation
problem is an ill-conditioned problem. This means that if
the search algorithm increases the yield and yet at the
same time decreases the maximum growth rate with a
certain fraction, the effect on the model output could be
cancelled out. The result is that many combinations of
parameter values for the yield and maximum growth rate
can have a similar effect on the model output. This is the
226
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
reason why a high correlation coefficient is obtained
after the parameter estimation has been performed. This
means that for a parameter to be uniquely identifiable,
100
their sensitivity functions should be unique and not
correlated with the sensitivity function of the other
parameters.
1.0
20
Ammonium
Nitrite
Nitrate
0
80
60
40
dNO3/dYAOB
dNO2/dYAOB
dNH4/dYAOB
0.5
-20
-40
0.0
-60
-0.5
20
1.2
2.4
Time (h)
3.6
-100
0.0
50
dAOB/dYAOB
-50
-100
3.6
0.5
0
0
0.0
3.6
NOO
0.0
-0.5
15
2.4
Time (h)
2.4
Time (h)
1.0
15
1.2
1.2
AOO
0
dO2/dYAOB
-1.0
0.0
3.6
20
Oxygen
-150
0.0
1.2
2.4
Time (h)
dNOB/dYAOB
0
0.0
-80
1.2
2.4
Time (h)
3.6
-1.0
0.0
1.2
2.4
Time (h)
3.6
Figure 5.9 Absolute sensitivity of the AOO yield on all the model outputs.
Another point of interest regarding these plots is that
the relative effect (that is, the magnitude of values on the
y-axis) of the parameters on ammonium, oxygen and
nitrite is quite similar. This means that all three of these
variables are equally relevant and important for
estimating these parameters.
Step 3. Parameter-significance ranking.
In this step the significance of parameters is ranked by
summarizing the non-dimensional sensitivity functions
of the parameters to model outputs using the δmsqr
measure. The results are shown in Figure 5.11.
The results show that the decay rate of AOO has
almost zero effect on all three of the measured variables
(ammonium, nitrite and oxygen) and therefore cannot be
estimated. This is known from process engineering and
for this reason, short-term batch tests are not used to
determine decay constants. This result therefore is a
confirmation of the correctness of the sensitivity
analysis. With regards to the maximum growth rate and
yield, these parameters are equally important followed
by the affinity constant for oxygen and ammonium. This
indicates that at least four parameters can potentially be
estimated from the data set.
DATA HANDLING AND PARAMETER ESTIMATION
227
YAOO
µmaxAOO
15
bAOO
KoAOO
KsAOO
1.0
15
Ammonium
Nitrite
10
Nitrate
10
0.5
0
-5
dy/dθ
5
dy/dθ
dy/dθ
5
0
0.0
-5
-0.5
-10
-10
-15
0.0
1.2
2.4
Time (h)
-15
0.0
3.6
20
0.0
3.6
3
AOO
10
-10
1
-1
2.4
Time (h)
3.6
0.0
-0.5
0
1.2
3.6
0.5
dy/dθ
dy/dθ
0
2.4
Time (h)
NOO
2
-20
0.0
1.2
1.0
Oxygen
dy/dθ
-1.0
1.2
2.4
Time (h)
-1.0
0
1.2
2.4
Time (h)
3.6
0.0
1.2
2.4
Time (h)
3.6
Figure 5.10 Relative sensitivity functions of the AOO parameters on the model outputs.
Step 5. Identifiability analysis.
dtm
= sqrt(det(asm))^(1/(i*2));
%determinant index
In this step, normalized sensitivity functions are used to
assess which parameter subsets have a small collinearity
index. The collinearity index is a measure of how two
sensitivity functions are aligned together, therefore
implying linear dependency.
for i = 2:subset
combos = combnk(set,i); % all possible
parameter combinations of different subset size
(2,3,4...)
for j=1:n
tempn
= snormy(:,combos(j,:)) ;
tempa
= say(:,combos(j,:)) ;
nsm
= tempn'*tempn;
% normalized
sensitivity matrix
asm
= tempa'*tempa;
sensitivity matrix, fim
% absolute
col
= 1/sqrt(min(eig(nsm))); %
collinearity index
subs(j,:)
= [k i col dtm] ;
end
end
The identifiability analysis indicates that there are 26
different combinations of the parameter subsets that can
potentially be used for parameter estimation using the
ammonium, nitrite and oxygen measurements (Table
5.8). The collinearity index value was changed from 1.2
to 53 and in general tends to increase for larger parameter
subset sizes. The parameter subset K#21 is the one used
in the parameter estimation above (see examples 5.3 and
5.4). This subset has a collinearity index of 45, which is
far higher than typically considered threshold values of
5-15 for a subset to be considered practically identifiable
228
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
(Brun et al., 2002; Sin et al., 2010). As shown here, the
analysis would have diagnosed the issue before
performing the parameter estimation (PE) and this would
AOO
µmax
YAOO
Ammonium
δmsqr
5
have indicated that this subset was not suitable for the
estimation.
KoAOO
KsAOO
0
µAOO
max
YAOO
Nitrite
δmsqr
5
bAOO
KoAOO
KsAOO
bAOO
0
AOO
µmax
YAOO
Oxygen
δmsqr
4
KoAOO
KsAOO
bAOO
0
4
AOO
δmsqr
YAOO
AOO
µmax
bAOO
KoAOO
KsAOO
4
5
0
1
2
3
Importance rank
Figure 5.2 Significance ranking of the AOO parameters with respect to the model outputs.
However, given that the sensitivity of bAOO was not
influential on the outputs (see Step 3), any subset
containing this parameter would not be recommended for
parameter estimation. Nevertheless there remain many
subsets that meet a threshold of 5-15 for γK that can be
considered for the parameter estimation problem. The
parameter subsets shaded in Table 5.8 meet these
identifiability criteria, and therefore can be used for
parameter estimation. The best practice is to start with
the parameter subset with the largest size (of parameters)
and lowest γK. Taking these considerations of the
sensitivity and collinearity index of the parameter
subsets into account helps to avoid the ill-conditioned
parameter estimation problem and to improve the quality
of the parameter estimates.
Example 5.6 Estimate the model prediction uncertainty of the
nitrification model – the Monte Carlo method
In this example, we wish to propagate the parameter
uncertainties resulting from parameter estimation (e.g.
Example 5.3 and Example 5.4) to model output
uncertainty using the Monte Carlo method.
For the uncertainty analysis, the problem is defined
as follows: (i) only the uncertainty in the estimated AOO
parameters is considered, (ii) the experimental
conditions of batch test 1 are taken in account (Table
5.3), and (iii) the model in Table 5.1 is used to describe
the system and nominal parameter values in Table 5.2.
DATA HANDLING AND PARAMETER ESTIMATION
229
Table 5.8 The collinearity index calculation for all the parameter combinations.
Subset K
Subset size
Parameter combination
1
2
Ko,AOO
bAOO
1.32
2
2
Ks,AOO
bAOO
1.26
3
2
Ks,AOO
Ko,AOO
2.09
AOO
max
γK
4
2
μ
bAOO
1.30
5
2
μmaxAOO
Ko,AOO
13.92
6
2
μmaxAOO
Ks,AOO
2.03
7
2
YAOO
bAOO
1.28
8
2
YAOO
Ko,AOO
12.55
9
2
YAOO
Ks,AOO
2.02
AOO
max
10
2
YAOO
μ
11
3
Ks,AOO
Ko,AOO
AOO
max
42.93
bAOO
2.10
12
3
μ
Ko,AOO
bAOO
14.05
13
3
μmaxAOO
Ks,AOO
bAOO
2.03
14
3
μ
AOO
max
Ks,AOO
Ko,AOO
14.23
15
3
YAOO
Ko,AOO
bAOO
13.09
16
3
YAOO
Ks,AOO
bAOO
2.02
17
3
YAOO
Ks,AOO
Ko,AOO
12.89
18
3
YAOO
μmaxAOO
bAOO
51.25
YAOO
μ
AOO
max
Ko,AOO
45.87
AOO
max
Ks,AOO
19
3
20
3
YAOO
μ
21
4
YAOO
μmaxAOO
Ks,AOO
Ko,AOO
45.91
43.37
22
4
YAOO
μmaxAOO
Ks,AOO
bAOO
51.25
AOO
max
Ko,AOO
bAOO
53.01
23
4
YAOO
μ
24
4
YAOO
Ks,AOO
Ko,AOO
bAOO
13.30
25
4
μmaxAOO
Ks,AOO
Ko,AOO
bAOO
14.30
Ks,AOO
Ko,AOO
26
5
YAOO
μ
AOO
max
Step 1. Input uncertainty definition.
As defined in the above problem definition, only the
uncertainties in the estimated AOO parameters are taken
into account:
bAOO
53.07
can be verified by calculating the empirical density
function for each parameter using the parameter
estimates matrix (θ50x4) and shown in Figure 5.12.
figure
labels=['\theta_1';'\theta_2';'\theta_3';'\theta
•
θinput = [YAOO µmax
AOO
Ks,AOO Ko,AOO].
Mean and standard deviation estimates are taken as
obtained from the bootstrap method together with their
correlation matrix (Table 5.7). Further it is assumed that
these parameters follow a normal distribution or
multivariate normal distribution since they have a
covariance matrix and are correlated. This assumption
_4']; %or better the name of parameter
for i=1:4
subplot(2,2,i)
[f xi]=ksdensity(pmin(:,i));
plot(xi,f)
xlabel(labels(i,:),'FontSize',fs,'FontWeight','b
old')
end
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
40
4
30
3
Counts
Counts
230
20
10
0
0.10
2
1
0.12
0.14
0.16
0
1.0
0.18
1.2
1.4
Distribution of θ1
1.6
1.8
Distribution of θ2
20
6
15
Counts
Counts
4
10
2
5
0
0.40
0
0.45
0.50
0.55
0.60
Distribution of θ3
0.4
0.8
0.6
1.0
Distribution of θ4
Figure 5.12 Empirical probability density estimates for the AOO parameters as obtained by the bootstrap method.
Step 2. Sampling from the input space
Since the input parameters have a known covariance
matrix, any sampling technique must take this into
account. In this example, since the parameters are
defined to follow a normal distribution, the input
uncertainty space is represented by a multivariate normal
distribution. A random sampling technique is used to
sample from this space:
%% do random sampling
N= 100; %% sampling number
mu=mean(pmin); %% mean values of parameters
sigma=cov(pmin); %% covariance matrix (includes
stand dev and correlation information)
X = mvnrnd(mu,sigma,N); % sample parameter space
using multivariate random sampling
The output from this step is a sampling matrix, XNxm,
where N is the sampling number and m is the number of
inputs. The sampled values can be viewed using a matrix
plot as in Figure 5.13. In this figure, which is a matrix
plot, the diagonal subplots are the histogram of the
parameter values while the non-diagonal subplots show
the sampled values of the two pairs of parameters. In this
case the most important observations are that (i) the
parameter input space is sampled randomly and (ii) the
parameter correlation structure is preserved in the
sampled values.
Step 3. Perform the Monte Carlo simulations.
In this step, N model simulations are performed using the
sampling matrix from Step 2 (XNxm) and the model
outputs are recorded in a matrix form to be processed in
the next step.
%%step 2 perform monte carlo simulations for
each parameter value
% Solution of the model
initcond;options=odeset('RelTol',1e7,'AbsTol',1e-8);
for i=1:nboot
disp(['the iteration number is :
',num2str(i)])
par(idx) = X(i,:) ; %read a sample from
sampling matrix
[t,y1] = ode45(@nitmod,td,x0,options,par); ;
%solve the model
y(:,:,i)=y1; %record the outputs
end
DATA HANDLING AND PARAMETER ESTIMATION
231
YAOO
0.20
0.15
0.10
AOO
µmax
2.0
1.5
1.0
KSAOO
0.6
0.5
0.4
KoAOO
1.0
0.8
0.6
0.1
0.15
0.2
1.0
1.5
2.0
0.4
0.5
AOO
µmax
YAOO
0.6
0.6
0.8
KsAOO
1.0
KoAOO
20
20
15
15
Nitrite (mg N L-1)
Ammonium (mg N L-1)
Figure 5.13 Plotting of the sampling matrix of the input space, XNxm – the multivariate random sampling technique with a known covariance matrix.
10
5
0
10
5
0
0.0
1.2
2.4
3.6
0.0
1.2
Time (h)
8
3.6
2.4
3.6
79
AOO (mg COD L-1)
Oxygen (mg O2 L-1)
2.4
Time (h)
6
4
78
77
76
2
75
0.0
1.2
2.4
3.6
Time (h)
Figure 5.14 Monte Carlo simulations (N = 100) of the model outputs.
0.0
1.2
Time (h)
232
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Step 4. Review and analyse the results.
In this step, the outputs are plotted and the results are
reviewed. In Figure 5.14, Monte Carlo simulation results
are plotted for four model outputs.
This means that while there is uncertainty in the
parameter estimates themselves, when the estimated
parameter subset is used together with its covariance
matrix, the uncertainty in the model prediction is low.
For any application of these model parameters they
should be used together as a set, rather than individually.
20
20
15
15
Nitrite (mg N L-1)
Ammonium (mg N L-1)
As shown in Figure 5.15, the mean, standard
deviation and percentiles (e.g. 95 %) can be calculated
from the output matrix. The results indicate that for the
sources of uncertainties being studied, the uncertainty in
the model outputs can be considered negligible. These
results are in agreement with the linear error propagation
results shown in Figure 5.5.
10
5
0
0.0
1.2
2.4
10
5
0
0.0
3.6
1.2
Time (h)
3.6
2.4
3.6
78
AOO (mg COD L-1)
Oxygen (mg O2 L-1)
8
6
4
2
0.0
2.4
Time (h)
1.2
2.4
77
76
75
0.0
3.6
Time (h)
1.2
Time (h)
Figure 5.15 Mean and 95 % percentile calculation of the model output uncertainty.
Another point to make is that the output uncertainty
evaluated depends on the input uncertainty defined as
well as the framing, e.g. initial conditions of the
experimental setup. For example, in the above example
what was not considered is the measurement uncertainty
or uncertainty due to other fixed parameters (decay) and
initial conditions (the initial concentration of autotrophic
bacteria). Therefore these results need to be interpreted
within the context where they are generated.
5.5 ADDITIONAL CONSIDERATIONS
Best practice in parameter estimation
In practice, while asymptotic theory assumption gives
reasonable results, there are often deviations from the
assumptions. In particular:
DATA HANDLING AND PARAMETER ESTIMATION
•
•
The measurement errors are often auto-correlated,
meaning that too many observations are redundant
and not independent (non-independently and
identically distributed (iid) random variables). This
tends to cause an underestimation of asymptotic
confidence intervals due to smaller sample variance,
σ2. A practical solution to this problem is to check the
autocorrelation function of the residuals and filter
them or perform subsampling such that
autocorrelation is decreased in the data set. The
parameter estimation can then be redone using the
subsample data set.
Parameter estimation algorithms may stop at local
minima, resulting in an incorrect linearization result
(the point at which the non-linear least squares are
linearized). To alleviate this issue, parameter
estimation needs to be performed several times with
either different initial guesses, different search
algorithms and/or an identifiability analysis.
Afterwards it is important to verify that the minimum
solution is consistent with different minimization
algorithms.
233
for the parameter values, is taken for granted. Hence a
proper analysis and definition of the parameter
estimation problem will always require a good
engineering judgment. For robust parameter estimation
in practice, due to the empirical/experiental nature of
parameter definition, the statistical methods (including
MLE estimates) should be treated within the
context/definition of the problem of interest.
With regards to bootstrap sampling, the most
important issue is whether or not the residuals are
representative of typical measurement error. For a more
detailed discussion of this issue, refer to Efron (1979).
Best practice in uncertainty analysis
When performing uncertainty analysis, the most
important issue is the framing and the corresponding
definition of the input uncertainty sources. Hence, the
outcome from an uncertainty analysis should not be
treated as absolute but dependent on the framing of the
analysis. A detailed discussion of these issues can be
found elsewhere (e.g. Sin et al., 2009; Sin et al., 2010).
Identifiability or ill-conditioning problem: Not all the
parameters can be estimated accurately. This can be
caused by a too large confidence interval compared to the
mean or optimized value of the parameter estimators.
The solution is to perform an identifiability analysis or
re-parameterisation of the model, so that a lower number
of parameters needs to be estimated.
Another important issue is the covariance matrix of
the parameters (or correlation matrix), which should be
obtained from a parameter estimation technique.
Assuming the correlation matrix is negligible may lead
to over or under estimation of the model output
uncertainty. Hence, in a sampling step the appropriate
correlation matrix should be defined for inputs (e.g.
parameters) considered for the analysis.
While we have robust and extensive statistical
theories and methods relevant for estimation of model
parameters as demonstrated above, the definition of the
parameter estimation problem itself, which is concerned
with stating what is the data available, what is the
candidate model structure, and what is the starting point
Regarding the sampling number, one needs to iterate
several times to see if the results differ from one iteration
to another. Since the models used for the parameter
estimation are relatively simple to solve numerically, it
is recommended to use a sufficiently high number of
iterations e.g. 250 or 500.
References
Brun, R., Kühni, M., Siegrist, H., Gujer, W., Reichert, P. (2002).
Practical identifiability of ASM2d parameters - systematic
selection and tuning of parameter subsets. Water Res. 36(16):
4113-4127.
Brun, R., Reichert, P., and Künsch, H. R. (2001). Practical
identifiability analysis of large environmental simulation
models. Water Resources Research, 37(4):1015-1030.
Bozkurt, H., Quaglia, A., Gernaey, K.V., Sin, G. (2015). A
mathematical programming framework for early stage design
of wastewater treatment plants. Environmental Modelling &
Software, 64: 164-176.
Dochain, D., Vanrolleghem, P.A. (2001). Dynamical Modelling
and Estimation in Wastewater Treatment Processes. London
UK: IWA Publishing.
Efron, B. (1979). Bootstrap methods: another look at the jackknife.
The Annals of Statistics, 7(1):1-26.
Gernaey, K.V., Jeppsson, U., Vanrolleghem, P.A., Copp, J.B.
(Eds.). (2014). Benchmarking of control strategies for
wastewater treatment plants. IWA Publishing.
Guisasola, A., Jubany, I., Baeza, J.A., Carrera, J., Lafuente, J.
(2005). Respirometric estimation of the oxygen affinity
constants for biological ammonium and nitrite oxidation.
Journal of Chemical Technology and Biotechnology, 80(4):
388-396.
234
Heijnen,
J.J.
(1999).
Bioenergetics
of
microbial
growth. Encyclopaedia of Bioprocess Technology.
Henze, M., Gujer, W., Mino, T., van Loosdrecht, M.C.M., (2000).
ASM2, ASM2d and ASM3. IWA Scientific and Technical
Report, 9. London UK.
Ljung L. (1999). System identification - Theory for the user. 2nd
edition. Prentice-Hall.
Mauricio-Iglesias, M., Vangsgaard, A.K., Gernaey, K.V., Smets,
B.F., Sin, G. (2015). A novel control strategy for single-stage
autotrophic nitrogen removal in SBR. Chemical Engineering
Journal, 260: 64-73.
Meijer, S.C.F., Van Der Spoel, H., Susanti, S., Heijnen, J.J., van
Loosdrecht, M.C.M. (2002). Error diagnostics and data
reconciliation for activated sludge modelling using mass
balances. Water Sci Tech. 45(6): 145-156.
Metropolis, N., Ulam, S. (1949). The Monte Carlo method. Journal
of the American Statistical Association, 44(247): 335-341.
Omlin, M. and Reichert, P. (1999). A comparison of techniques for
the estimation of model prediction uncertainty. Ecol. Model.,
115: 45-59.
Roels, J.A. (1980). Application of macroscopic principles to
microbial
metabolism. Biotechnology
and
Bioengineering, 22(12): 2457-2514.
Saltelli, A., Tarantola, S., and Campolongo, F. (2000). Sensitivity
analysis as an ingredient of modeling. Statistical Science,
15(4):377-395.
Seber G. and Wild C. (1989) Non-linear regression. Wiley, New
York.
Sin, G., Gernaey, K.V., Neumann, M.B., van Loosdrecht, M.C.M.,
Gujer, W. (2009). Uncertainty analysis in WWTP model
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
applications: a critical discussion using an example from
design. Water Res. 43(11): 2894-2906.
Sin, G., Gernaey, K.V., Neumann, M.B., van Loosdrecht, M.C.M.,
Gujer, W. (2011). Global sensitivity analysis in wastewater
treatment plant model applications: prioritizing sources of
uncertainty. Water Research, 45(2): 639-651.
Sin, G., de Pauw, D.J.W., Weijers, S., Vanrolleghem, P.A. (2008).
An efficient approach to automate the manual trial and error
calibration of activated sludge models. Biotechnology and
Bioengineering. 100(3): 516-528.
Sin, G., Meyer, A.S., Gernaey, K.V. (2010). Assessing reliability of
cellulose hydrolysis models to support biofuel process designidentifiability and uncertainty analysis. Computers &
Chemical Engineering, 34(9): 1385-1392.
Sin, G., Vanrolleghem, P.A. (2007). Extensions to modeling
aerobic carbon degradation using combined respirometric–
titrimetric measurements in view of activated sludge model
calibration. Water Res. 41(15): 3345-3358.
Vangsgaard, A.K., Mauricio-Iglesias, M., Gernaey, K.V., Sin, G.
(2014). Development of novel control strategies for singlestage autotrophic nitrogen removal: A process oriented
approach. Computers & Chemical Engineering, 66: 71-81.
van der Heijden, R.T.J.M., Romein, B., Heijnen, J.J., Hellinga, C.,
Luyben, K. (1994). Linear constraint relations in biochemical
reaction systems: II. Diagnosis and estimation of gross
errors. Biotechnology and bioengineering, 43(1): 11-20.
Villadsen, J., Nielsen, J., Lidén, G. (2011). Elemental and Redox
Balances. In Bioreaction Engineering Principles (pp. 63-118).
Springer, US.
6
SETTLING TESTS
Authors:
Reviewers:
Elena Torfs
Ingmar Nopens
Mari K.H. Winkler
Peter A. Vanrolleghem
Sophie Balemans
Ilse Y. Smets
Glen T. Daigger
Imre Takács
6.1 INTRODUCTION
Low
concentration
The settling behaviour of a suspension (e.g.
secondary sludge, raw wastewater or granular sludge) is
governed by its concentration and flocculation tendency
and can be classified into four regimes (Figure 6.1):
discrete non-flocculent settling (Class I), discrete
flocculent settling (Class II), zone settling or hindered
settling (Class III) and compressive settling (Class IV).
Dilution
Settling is an important process in several of the unit
operations in wastewater treatment plants (WWTPs). The
most commonly known of these unit processes are
primary settling tanks (PSTs), which are a treatment units
before the biological reactor, and secondary settling tanks
(SSTs), which are a clarification step prior to discharge
into a receiving water. Moreover, settling also plays an
important role in new technologies that are being
developed such as granular sludge reactors. Due to the
different nature of the settleable components (raw
wastewater, activated sludge, and granular sludge) and
the concentration at which these compounds occur, these
unit processes are characterised by distinctly different
settling behaviours. This section presents an overview of
the different settling regimes that a particle-liquid
suspension can undergo and relates these regimes to the
specific settling behaviour observed in SSTs, PSTs and
granular sludge reactors.
Clarification
Class I
Clarification
Class II
Zone settling
Class III
Compression
Class IV
High
concentration
Low:
particulate
High:
flocculent
Inter-particle cohesiveness
Figure 6.1 Settling regimes (Ekama et al., 1997).
At low concentrations (classes I and II), the particles
are completely dispersed, there is no physical contact
between them and the concentration is typically too
diluted for particles to influence each other’s settling
behaviour. Each particle settles at its own characteristic
terminal velocity, which depends on individual particle
properties such as shape, size, porosity and density. If
© 2016 Elena Torfs et al. Experimental Methods In Wastewater Treatment. Edited by M.C.M. van Loosdrecht, P.H. Nielsen, C.M. Lopez-Vazquez and D. Brdjanovic. ISBN: 9781780404745
(Hardback), ISBN: 9781780404752 (eBook). Published by IWA Publishing, London, UK.
236
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
these dilute particles show no tendency to flocculate (for
example, granular sludge), this regime is called discrete
non-flocculent settling (Class I). However, certain
suspensions (among which raw wastewater solids and
activated sludge flocs) have a natural tendency to
flocculate even at low concentrations (Ekama et al.,
1997). Through subsequent processes of collision and
cohesion, larger flocs are formed causing their settling
velocity to change over time. This regime is called
discrete flocculent settling (Class II). It is important to
note that during discrete flocculent settling, the formed
flocs will still settle at their own characteristic terminal
velocity. Hence, both discrete non-flocculent and discrete
flocculent settling undergo basically the same settling
dynamics. The difference lies in the fact that, for discrete
flocculent settling, an additional flocculation process is
occurring simultaneously with the settling process, which
alters the particles’ individual properties and
consequently their terminal settling velocity.
6.1. Hence, in an SST, different settling regimes occur
simultaneously at different locations throughout the tank.
Low concentrations in the upper regions of the SST
favour discrete (flocculent) settling whereas the
concentrations of the incoming sludge are typically in the
range for hindered settling, and sludge thickening inside
the sludge blanket is governed by compressive settling.
In contrast to this, a specific characteristic of the granular
sludge technology is the granules’ low tendency to
coagulate under reduced hydrodynamic shear (de Kreuk
and van Loosdrecht, 2004), thus positioning them on the
left side of Figure 6.1. This feature causes granular
sludge to undergo discrete (non-flocculent) settling at
concentrations where conventional activated sludge
undergoes hindered or compression settling. Finally,
PSTs are fed by incoming wastewater containing a
relatively low concentration of suspended solids (i.e. the
upper part Figure 6.1). Hence, the dominant settling
regime in these tanks is discrete settling (classes I and II).
The transition from discrete settling to hindered
settling (Class III) occurs if the solid concentration in the
tank exceeds a threshold concentration where the
particles no longer settle independently of one another.
As can be seen from Figure 6.1, this threshold
concentration depends on the flocculation state of the
sludge. For secondary sludge the transition typically
occurs at concentrations of 600 - 700 mg TSS L-1 whereas
for granular sludge the threshold concentration can go up
to 1,600-5,500 mg TSS L-1 (depending on the granulation
state) (Mancell-Egala et al., 2016). Above this threshold,
each particle is hindered by the other particles and the
inter-particle forces are sufficiently strong to drag each
particle along at the same velocity, irrespective of size
and density. In other words, the particles settle
collectively as a zone, and therefore this regime is also
called zone settling. In this regime, a distinct interface
between the clear supernatant and the subsiding particles
is formed. When the solids concentration further
increases above a critical concentration (5-10 g L-1), the
settling behaviour changes to compressive settling (Class
IV). The exact transition concentration depends once
more on the flocculation state of the particles (De Clercq
et al., 2008). At these elevated concentrations, the solids
come into physical contact with one another and are
subjected to compaction due to the weight of overlying
particles. The settling velocity will be much lower than
in the hindered settling regime.
As each unit process is characterised by its own
settling behaviour, different experimental methods are
required to assess their performance. This chapter
provides an overview of experimental methods to analyse
the settling and flocculation behaviour in secondary
settling tanks (Section 6.2 and 6.3), granular sludge
reactors (Section 6.4) and primary settling tanks (Section
6.5).
Activated sludge with a proper biological make-up
shows a natural tendency to flocculate and, depending on
its concentration, can cover the entire right side of Figure
6.2 MEASURING SLUDGE SETTLEABILITY
IN SSTs
To evaluate the performance of an SST, it is essential to
quantify the settling behaviour of the activated sludge in
the system. Batch settling experiments are an interesting
information source in this respect since they eliminate the
hydraulic influences of in- and outgoing flows on the
settling behaviour. Therefore, several methods aim to
determine the settling characteristics of the activated
sludge by measuring certain properties during the settling
of activated sludge in a batch reservoir.
Several types of measurements can be performed by
means of a batch settling test. These measurements range
from very simple experiments providing a rough
indication of the sample’s general settleability (Section
6.2.1) to more labour-intensive experiments that measure
the specific settling velocity (Section 6.2.2) or even
determine a relation between the settling velocity and the
sludge concentration (Section 6.2.3). Moreover, some
useful recommendations to consider when performing
SETTLING TESTS
batch settling experiments are provided in Section 6.2.4
and an overview of recent developments with respect to
these type of experiments can be found in Section 6.2.5.
6.2.1 Sludge settleability parameters
6.2.1.1 Goal and application
A number of parameters have been developed to obtain a
quantitative measure of the settleability of an activated
sludge sample. These Sludge Settleability Parameters
(SSPs) are based on the volume that sludge occupies after
a fixed period of settling. Among these, the Sludge
Volume Index (SVI) (Mohlman, 1934) is the most
known. A number of issues have been reported with the
SVI as a measure of sludge settleability (Dick and
Vesilind, 1969; Ekama et al., 1997) of which the most
important one is its dependency on the sludge
concentration. Particularly at higher concentrations,
measured SVI values can deviate significantly between
sludge concentrations (Dick and Vesilind, 1969).
Moreover, SVI measurements have been found to be
influenced by the dimensions of the settling cylinder.
These problems can be significantly reduced by
conducting the test under certain prescribed conditions.
Hence, a number of modifications have been proposed to
the standard SVI test in order to yield more consistent
information (Stobbe, 1964; White, 1976, 1975). Stobbe
(1964) proposed conducting the SVI test with diluted
sludge and called it the Diluted SVI (DSVI). White
(1975, 1976) proposed the Stirred Specific Volume Index
(SSVI3.5) where the sludge sample is stirred during
settlement. Although each of these SSPs are described in
more detail below, it is important to note that the SSVI3.5
is known to provide the most consistent results.
6.2.1.2 Equipment
a. A graduated (minimum resolution 50 mL) cylindrical
reservoir with a volume of 1 litre (for SVI and DSVI)
or with dimensions specified by White (1975) for
SSVI.
b. A digital timer displaying accuracy in seconds.
c. A sludge sample from either the recycle flow of the
SST or the feed flow into the SST. The latter can be
collected from the bioreactor or the splitter structure.
d. Effluent from the same WWTP (in case dilution is
needed).
e. Equipment for the Total Suspended Solids (TSS) test
(according to method 2540 D in APHA et al., 2012).
f. A stirrer for the SSVI test.
237
6.2.1.3 The Sludge Volume Index (SVI)
The Sludge Volume Index (SVI) (Mohlman, 1934) is
defined as the volume (in mL) occupied by 1 g of sludge
after 30 min settling in a 1 L unstirred cylinder.
• Protocol
1. Measure the concentration of the sludge sample with
a TSS test according to method 2540 D of Standard
Methods (APHA et al., 2012).
2. Fill a 1 L graduated cylinder with the sludge sample
and allow the sample to settle.
3. After 30 min of settling, read the volume occupied by
the sludge from the graduated cylinder (SV30 in mL
L-1).
4. Calculate the SVI from Eq. 6.1, with XTSS being the
measured concentration of the sample in g L-1:
SVI =
SV30
XTSS
Eq. 6.1
• Example
In this example an SVI test is performed with a sludge
sample from the bioreactor in the WWTP at
Destelbergen. The concentration of the sample is
measured at 2.93 g L-1.
A graduated cylinder is filled with the sludge sample
and the sludge is allowed to settle. After 30 min of
settling, the sludge occupies a volume of 290 mL (SV30).
Hence the sample has an SVI of:
SVI =
290 mL L-1
= 99 mL g-1
2.93 g L-1
This result indicates a sludge with good settling
properties. Typical SVI values for AS can be found
between 50-400 mL g-1 where 50 mL g-1 indicates a
sample with very good settleability and 400 mL g-1 a
sample with poor settling properties.
6.2.1.4 The Diluted Sludge Volume Index (DSVI)
The Diluted Sludge Volume Index: DSVI (Stobbe, 1964)
differs from the standard SVI by performing an
additional dilution step prior to settling. The sludge is
hereby diluted with effluent until the settled volume after
30 min is between 150 ml L-1 and 250 ml L-1. Note that
all the dilutions must be made with effluent (before
chemical disinfection) from the plant where the sludge is
238
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
obtained to reduce the possibility of foreign substances
affecting the settling behaviour.
This value can then be used in Eq. 6.1 to calculate the
SSVI3.5 with XTSS = 3.5 g L-1.
• Protocol
1. Dilute the sludge sample with effluent until the
settled volume after 30 min is between 150 mL L-1
and 250 mL L-1.
2. Perform steps 1-3 from the standard SVI test.
3. Calculate the DSVI from Eq. 6.1, with XTSS being the
concentration of the diluted sample in g L-1.
Although the SSVI3.5 is not as easily executed as the
DSVI due to the specified stirring equipment required, it
does provide the most consistent results (Ekama et al.,
1997; Lee et al., 1983).
The advantage of the DSVI lies in its insensitivity to
the sludge concentration, allowing for consistent
comparison of sludge settleability between different
activated sludge plants.
6.2.1.5 The Stirred Specific Volume Index (SSVI3.5)
The Stirred Specific Volume Index (SSVI3.5) was
presented by White (1975, 1976) who found that stirring
the sample during settling reduces wall effects, short
circuiting and bridge formation effects, thereby creating
conditions more closely related to those prevailing in the
sludge blanket in SSTs.
The SSVI3.5 is determined by performing an SVI test
at a specific concentration of 3.5 g L-1 while the sludge is
gently stirred at a speed of about 1 rpm. To determine the
SSVI3.5, the sludge concentration is measured with a TSS
test and subsequently diluted with effluent to a
concentration of 3.5 g L-1. In some cases further
concentration of the sample may be necessary if the plant
is operating at MLSS values below 3.5 g L-1 and sampling
from the return activated sludge flow is not possible.
Compared to the DSVI, the SSVI3.5 has the additional
advantage that it not only overcomes the concentration
dependency but is shown to be relatively insensitive to
the dimensions of the settling column, provided that it is
not smaller than the dimensions specified by White
(1976), i.e. a depth to diameter ratio of between 5:1 and
6:1, and a volume of more than 4 L. As the measurement
is performed in a larger reservoir, the SSVI3.5 cannot be
directly calculated from Eq. 6.1 but the volume of the
particular column has to be reduced to an equivalent 1 L
column. This is done by expressing the settled sludge
volume at 30 min as a fraction of the column volume (fsv)
and multiplying this fraction by 1,000 mL to obtain the
equivalent l L stirred settled volume SSV30 (Eq. 6.2).
SSV30 = fsv · 1,000
Eq. 6.2
• Protocol
1. Measure the concentration of the sludge sample with
a TSS test according to method 2540 D of Standard
Methods (APHA et al., 2012).
2. Dilute the sample with effluent to a concentration of
3.5 g L-1.
3. Fill a graduated cylinder with the minimum
dimensions specified by White (1976).
4. After 30 min of settling during which the sample is
stirred at a speed of 1 rpm, read off the volume
occupied by the sludge from the graduated cylinder
and calculate the SSV30 from Eq. 6.2.
5. Calculate the SSVI3.5 from Eq. 6.1 with SV30 = SSV30
and XTSS = 3.5 g L-1.
6.2.2 The batch settling curve and hindered
settling velocity
6.2.2.1 Goal and application
The sludge settleability parameters presented above
provide a low level measurement of the general
settleability. However, it should be stressed that they
represent only a momentary recording of the settling
behaviour. In reality, the volume of a sludge sample after
30 min of settling will depend on both its hindered
settling and compression behaviour, which are both
influenced by a number of factors such as the
composition of the activated sludge (for example, the
population of filamentous organisms), floc size
distributions, surface properties, rheology, etc.
Consequently, two sludge samples with different settling
behaviour can result in similar values for the sludge
settleability parameters.
More detailed information on the settling behaviour
of a sludge sample can be obtained from a batch settling
curve which makes it possible to investigate the settling
behaviour of sludge at different settling times.
Batch settling curves can serve different purposes.
They can be used either qualitatively to determine
operational or seasonal trends in the settling behaviour or
SETTLING TESTS
they can be used quantitatively to determine the SST’s
capacity limit. In the former, a simple graduated cylinder
can be used (for example, the cylindrical reservoir in
Figure 6.2). In the latter, the selection of an appropriate
settling reservoir to avoid wall effects during the test is
imperative. More information on the optimal shape and
size of batch settling reservoirs is provided in Section
6.2.4.1.
239
3. Start the timer immediately after filling the column.
4. Measure the sludge water interface at the following
time intervals: 0, 0.5, 1, 2, 3, 4, 5, 10, 15, 20, 30 and
45 min.
6.2.2.2 Equipment
For this test the following equipment is needed:
a. A graduated (minimum resolution 50 mL) cylindrical
reservoir.
b. A digital timer displaying accuracy in seconds.
c. A sludge sample from either the recycle flow of the
SST or the feed flow into the SST. The latter can be
collected from the bioreactor or the splitter structure.
d. Equipment for a TSS test (APHA et al., 2012).
e. Stirring equipment if the results are to be used for
quantitative analysis of the SST’s capacity.
6.2.2.3 Experimental procedure
To measure a batch settling curve, a reservoir is filled
with a sludge sample and a timer is started to keep track
of the duration of the experiment. The sludge is allowed
to settle and the position of the suspension-liquid
interface is measured at different time intervals. This
methodology is illustrated in Figure 6.2 where the
position of the suspension-liquid interface is indicated by
the red arrow. Recording the height of the suspensionliquid interface at several time intervals results in a curve
with the evolution of the sludge blanket height over time.
Standard measurement times for a batch settling curve
are 0, 0.5, 1, 2, 3, 4, 5, 10, 15, 20, 30 and 45 min but these
can be adapted depending on the settling dynamics of a
specific sludge sample (for more information see Section
6.2.4). At the start of the test, the suspension-liquid
interface is typically measured more frequently, as the
sludge is settling at a relatively fast pace. Later in the test,
the frequency of the measurements is decreased, because
the interface is moving more slowly.
• Protocol
1. Homogenize the sludge sample. Do not shake the
sample vigorously as this will disturb the sludge and
alter its settling properties.
2. Fill the cylindrical reservoir with the sample. Pour
gently and in a steady flow so as not to disturb the
sludge too much nor to allow it to settle again in the
container.
Figure 6.2 Photograph of the batch settling column at different settling
times, indicating the suspension-liquid interface (photo: E. Torfs).
• Example
A batch settling curve is measured with a sludge sample
from the bioreactor in the WWTP at Destelbergen. The
concentration of the sample is measured as 2.93 g L-1.
The measured heights of the suspension-liquid interface
(i.e. the Sludge Blanket Height: SBH) at different settling
times are provided in Table 6.2 and the batch curve is
shown in Figure 6.3.
240
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 6.2 Measured sludge blanket height during a single batch settling
test.
SBH (m)
0.250
0.247
0.244
0.238
0.206
0.184
0.169
0.122
0.106
0.098
0.088
0.081
Log phase
(a)
Height of interface
Time (min)
0
0.5
1
2
3
4
5
10
15
20
30
45
column may differ from the settling behaviour at the
interface depending on the local concentrations.
Zone settling phase
(b)
Transition phase
(c)
Compression phase
0.25
Time
SBH (m)
0.20
Figure 6.4 Evolution of the sludge blanket height over time, indicating the
four phases (Rushton et al., 2000).
0.15
0.10
Figure 6.5 represents the distribution of settling
regions over the depth of the column at different times
during a batch settling experiment.
0.05
0.00
0.0
0.2
0.4
0.6
0.8
Time (h)
A
Figure 6.3 Measured batch settling curve.
A
A
6.2.2.4 Interpreting a batch settling curve
Typically, four different phases can be observed in a
batch settling curve. Each phase marks a change in the
settling behaviour at the suspension-liquid interface.
Figure 6.4 shows the evolution of the sludge blanket
height over time during a batch settling test, indicating
the four phases. It is important to note that a batch settling
curve only provides information on the settling behaviour
at the sludge-water interface. At any specific time, the
settling behaviour at different depths throughout the
B
B
B
C
C
D
D
D
Figure 6.5 Chronological process of a batch settling test (Ekama et al.,
1997).
SETTLING TESTS
From the beginning of the test up to point (a) (Figure
6.4), the suspension-liquid interface is in the lag phase.
In this phase, the activated sludge needs to recover from
disturbances due to turbulence caused by the filling of the
batch column.
During the hindered settling phase or zone settling
phase which starts at point (a) and ends at point (b)
(Figure 6.4), the interface is located in the hindered
settling region (region B). This phase is characterised by
a distinct linear decline in the batch curve. An
equilibrium between the gravitational forces causing the
particles to settle and the hydraulic friction forces
resisting this motion results in the same settling velocity
for all the particles in the region. If the column is not
stirred, then the velocity at which the interface moves
downward is called the hindered settling velocity vhs at
the inlet solids concentration. If the dimensions and
conditions for the batch test were set so as to avoid wall
effects (Section 6.2.4.1), the measured settling velocity
corresponds with the zone settling velocity in an actual
SST.
The transition phase starts (point (b)) when the sludge
blanket reaches the transition layer (region C). The
transition layer is a layer of constant thickness and is
formed by particles coming from the decreasing hindered
settling layer and particles coming from the increasing
compression layer. Although during this phase the same
characteristics exist as in the zone settling regime, the
settling velocity decreases because the concentration
gradient increases with depth. The transition phase ends
when the sludge blanket reaches the compression layer.
The last phase starts at point (c) and is called the
compression phase. The time at which the compression
phase starts, called the compression point, is difficult to
identify. During the compression phase the particles
undergo compaction, thus creating an increasing
concentration gradient as well as a decreasing settling
velocity.
6.2.2.5 Measuring the hindered settling velocity
At moderate sludge concentrations (between approx. 1 g
L-1 and 6 g L-1), sludge will initially settle according to
the zone or hindered settling regime. The slope of the
linear part of a batch settling curve corresponds to the
hindered settling velocity vhs.
• Example
The hindered settling velocity for the data in Figure 6.3
is computed by determining the steepest slope between
three consecutive data points (which can be performed in
any software). The procedure is illustrated in Figure 6.6
and results in a settling velocity of 1.374 m h-1.
0.25
0.20
SBH (m)
Almost immediately after the start-up of the
experiment, four regions are formed at increasing depth.
The top region (region A) consists of supernatant. Below
region A, regions B, C and D are formed where
respectively zone settling, transition settling and
compression settling take place (Ekama et al., 1997). The
position of the sludge/water interface is indicated with a
red arrow. Hence, the phases that are recorded in a batch
settling curve occur as the suspension-liquid interface
passes through these different settling regions.
241
0.15
0.10
0.05
0.00
0.00
0.2
0.4
0.6
0.8
Time (h)
Figure 6.6 Calculation of the steepest slopes from a batch settling curve at
a concentration of 2.93 g L-1.
6.2.3 vhs-X relation
6.2.3.1 Goal and application
The concentration in the hindered settling region is
uniform and equal to the initial solids concentration of
the batch. By calculating the slopes of the linear part of
the batch curves for different initial concentrations, the
hindered settling velocity can be determined as a function
of the solids concentration. The relation between
hindered settling velocity and concentration is of
particular importance for the design of SSTs as it governs
the determination of the limiting flux and thus the SST’s
surface area. As stated previously, in order to use the
vhs-X relation for quantitative calculations such as the
determination of the SST’s surface area, the dimensions
242
6.2.3.2 Equipment
For this test the following equipment is needed:
a. A graduated (minimum resolution 50 mL) cylindrical
reservoir.
b. A digital timer displaying accuracy in seconds.
c. A sludge sample from the recycle flow of the SST.
d. Effluent from the same WWTP (for dilution).
e. Equipment for the TSS test (APHA et al., 2012).
f. Stirring equipment if the results are to be used for
quantitative analysis of the SST’s capacity.
6.2.3.3 Experimental procedure
To obtain different initial concentrations for the batch
experiments, a sludge sample from the recycle flow of
the SST is diluted with effluent from the same WWTP.
Hence, a dilution series is made with respectively 100,
80, 60, 50, 40 and 20 % of sludge. For example, a 40 %
dilution consists of 0.8 L of sludge from the recycle flow
and 1.2 L of effluent. For each dilution a batch curve is
measured according to the step-wise protocol from
Section 6.2.2.3 and the hindered settling velocity is
calculated from the slope of the linear part of the batch
curve.
In order to obtain reliable results for the vhs-X
relation, it is important to have an accurate measure of
the initial concentration in each experiment. The
concentrations in the dilution series are often determined
by measuring the concentration in the recycle flow and
then calculating the concentration of the dilution
assuming the effluent concentration is negligible.
However, this procedure is prone to errors if the recycle
flow sample is not fully mixed at any time during the
filling of the batch. A more reliable approach is to
measure the TSS of each dilution experiment separately.
This can be done by mixing up the content of the batch
reservoir at the end of each experiment and subsequently
taking a sample for the TSS measurement. This approach
requires some additional work as more TSS tests need to
be performed but it ensures a reliable measurement of the
diluted concentration.
• Protocol
1. Perform Step 1 from the protocol to measure the
batch curves with a sludge sample from the recycle
flow.
2. Combine a certain volume of the sludge sample with
the effluent until the required dilution is obtained.
3. Perform steps 2 to 4 from the protocol to measure the
batch curves.
4. After 45 min of settling, homogenise the sample in
the cylindrical reservoir again and take a sample to
determine the sludge concentration with a TSS test.
• Example
Samples were collected from the recycle flow and the
effluent at the WWTP in Destelbergen (Belgium). The
sludge/water interface during settling was measured for
different initial concentrations (Table 6.3). The resulting
settling curves are shown in Figure 6.7.
Table 6.3 Measured sludge blanket height (in m) during batch settling
tests at different initial concentrations.
Time 1.37 g L-1
0
0.248
0.5
0.243
1
0.215
2
0.107
3
0.074
4
0.064
5
0.059
10
0.046
15
0.041
20
0.038
30
0.033
45
0.031
2.37 g L-1
0.248
0.244
0.236
0.198
0.163
0.144
0.130
0.102
0.091
0.083
0.073
0.064
3.42 g L-1 4.10 g L-1
0.248
0.248
0.246
0.248
0.241
0.247
0.214
0.244
0.186
0.242
0.165
0.239
0.149
0.234
0.115
0.195
0.102
0.172
0.092
0.156
0.083
0.132
0.074
0.114
5.46 g L-1
0.248
0.247
0.246
0.245
0.243
0.243
0.241
0.234
0.227
0.219
0.205
0.182
6.83 g L-1
0.248
0.248
0.248
0.248
0.247
0.246
0.245
0.241
0.239
0.236
0.231
0.223
0.25
6.83 g/L
0.20
5.46 g/L
SBH (m)
and conditions of the batch settling test need to be set
according to the specifications given in Section 6.2.4.1.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
0.15
4.10 g/L
0.10
3.42 g/L
2.73 g/L
0.05
1.37 g/L
0.00
0.0
0.2
0.4
0.6
Time (h)
Figure 6.7 Batch settling curves at different initial concentrations.
0.8
SETTLING TESTS
243
The hindered settling velocities for the data in Figure
6.7 are computed by determining the steepest slope
between three consecutive data points (Figure 6.8A). The
resulting velocities are presented in Figure 6.8 B and
Table 6.4.
slope and the validity of these curves to measure hindered
settling may be questioned. More information can be
found in Section 6.2.4.3.
Table 6.4 Measured hindered settling velocities at different initial
concentrations.
Concentration (g L-1)
vhs (m h-1)
1.37
2.73
3.42
4.10
5.46
6.83
4.39
2.01
1.53
0.46
0.09
0.05
A)
0.25
6.83 g/L
0.2
SBH (m)
5.46 g/L
0.15
4.10 g/L
0.1
6.2.3.4 Determination of the zone settling
parameters
3.42 g/L
2.73 g/L
0.05
1.37 g/L
0.00
0.0
0.2
0.4
0.6
0.8
Time (h)
5
Hindered settling velocity (m h-1)
B)
4
3
Mathematically, the relation between the sludge
concentration and the zone settling velocity can be
described by an exponential decaying function (Eq. 6.3)
(Vesilind, 1968). In this equation vhs represents the
hindered settling velocity of the sludge, V0 the maximum
settling velocity, XTSS the solids concentration and rV a
model parameter. The parameters V0 and rV in this
function provide information on the sludge settleability
and are frequently used in SST design procedures. More
information on design procedures can be found in Ekama
et al. (1997).
vhs (X) = V0 ∙ e-rV ∙XTSS
2
1
0
0
2
4
6
8
10
Sludge concentration (g/L)
Figure 6.8 (A) Batch settling experiments at different initial solids
concentration indicating the maximum slope for each curve. (B) The
maximal slope represents a measurement of the hindered settling velocity.
The hindered settling velocity slows down at higher
concentrations because the settling particles will be
increasingly hindered by surrounding particles. Note that
for the concentrations 5.46 g L-1 and 6.83 g L-1, it
becomes increasingly difficult to determine the steepest
Eq. 6.3
The parameters V0 and rV can be estimated from the
experimental data by minimising the Sum of Squared
Errors (SSE) in Eq. 6.4 In this equation N is the number
of data points, vhs,i the measured hindered settling
velocity at concentration i, and ṽhs,i is the corresponding
prediction by the function of Vesilind (1968) for a
particular parameter set [V0 rV].
N
vhs,i ‒ ṽhs,i (V0 ,rV )
SSE =
2
Eq. 6.4
i=1
An estimation of the zone settling parameters and
calculation of the confidence intervals for these estimates
can be performed as explained in Chapter 5. Minimising
the SSE from Eq. 6.4 will give more weight to the fit of
high settling velocities (i.e. at low concentrations). In
order to give equal weight to all the measured settling
velocities, a logarithmic fit can be performed.
244
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
• Example
Table 6.5 provides the initial parameter estimates, the
optimal parameters after optimisation and the 95 %
confidence intervals for the estimated parameters for the
data in Table 6.4. The simulation results of calibrated
functions vs. the experimental data points are shown in
Figure 6.9.
Table 6.5 Initial values, optimal values and confidence intervals of the
estimated parameters for the settling function.
Parameter
V0 (m h-1)
rV (L g-1)
Initial value
9.647
0.488
Optimal value
10.608
0.634
Confidence interval
± 1.265
± 0.038
12
Data
Vesilind function
Zone settling velocity (m h-1)
10
8
6
4
6.2.3.5 Calibration by empirical relations based on
SSPs
As the measurement of batch settling curves is much
more time-consuming than the measurements of simple
SSPs, several empirical equations have been developed
that relate the settling parameters V0 and rV to simple
measurements of SSPs (Härtel and Pöpel, 1992;
Koopman and Cadee, 1983; Pitman, 1984, Daigger and
Roper, 1985). Examples of such empirical equations and
the resulting parameter estimates are given in Table 6.6.
Figure 6.10 shows that the settling parameters
calculated from the empirical equations are not able to
accurately describe the measured data from Table 6.4.
This could be expected as sludges with a similar SVI may
show a different settling behaviour dependent on the
sludge properties. Moreover, when using the empirical
relations, two parameters are estimated based on only one
data point. The SSPs thus provide insufficient
information to describe the settling behaviour at different
sludge concentrations. For these reasons, the use of
empirical relations based on SSPs is not an accurate
method to estimate the hindered settling parameters of
the settling functions and should be avoided.
2
0
0
2
4
6
8
10
Sludge concentration (g L-1)
Figure 6.9 Settling velocity as a function of the solids concentration. The
circles represent measured settling velocities and the line the calculated
settling velocities after calibration of the function by Vesilind (1968).
Table 6.6 Estimated values for V0 (m h-1) and rV (L g-1) by empirical relations based on SSPs.
Reference
Härtel and Pöpel (1992)
Equations
Parameter values
-0.0113 SVI
V0 = 9.647
Eq. 6.5
+ 1.043
rV = 0.488
Eq. 6.6
ln(V0 ) = 2.605 ‒ 0.00365 DSVI
V0 = 9.993
Eq. 6.7
rV = 0.249 + 0.002191 DSVI
rV = 0.431
Eq. 6.8
V0
= 67.9 e-0.016 SSVI3.5
rV
V0 = 5.669
Eq. 6.9
rV = 0.446
Eq. 6.10
V0 = 17.4 e
-0.00581 SVI
rV = ‒0.9834 e
Koopman and Cadee (1983)
Pitman (1984)
Equation nr.
rV = 0.88 ‒ 0.393 log
V0
rV
SETTLING TESTS
245
10
Data
Härtel and Pöpel (1992)
Koopman and Cadee (1983)
Pitman (1984)
Settling velocity (m h-1)
8
6
4
2
0
0
2
4
6
8
10
Solids concentration (g L-1)
Figure 6.10 Settling velocity as a function of the solids concentration. The
circles represent the measured settling velocities from the batch settling tests
and the lines represent the settling velocities calculated by the function of
Vesilind (1986) with parameter values based on empirical equations.
6.2.4.3 Concentration range
Hindered
settling
occurs
typically
between
concentrations of 1 g L-1 and 6 g L-1. However, these
limits are dependent on the flocculation state of the
sludge and may be different between WWTPs.
Therefore, there should always be a visual check on
whether the recorded batch settling curves are within the
hindered settling region. At low initial concentrations, it
becomes increasingly difficult to track the solid-liquid
interface as the sludge enters the discrete settling regime.
If no distinct interface can be observed, this
concentration should not be considered in the analysis.
On the other hand, at high concentrations, sludge starts to
undergo compression. If the initial concentration is high
enough for compression to occur at the very beginning of
the experiment, the batch curve will no longer show a
clear linear descent. If no linear decrease in the batch
curve can be seen, the high concentration should also not
be considered in the analysis.
6.2.4.4 Measurement frequency
6.2.4 Recommendations for performing
batch settling tests
6.2.4.1 Shape and size of the batch reservoir
It is recommended to use a cylindrical reservoir for batch
settling tests. In conical reservoirs, such as Imhoff cones,
no zone settling region can be measured because the
reducing width of the cross section will inevitably cause
a concentration gradient. When the goal of the
measurements is to use the zone settling parameters for a
qualitative analysis of the limiting flux and the SST’s
surface area, care should be taken to avoid any influence
of the settling reservoir on the settling behaviour (socalled wall effects). This can be achieved by performing
the batch test in a column of at least 100 mm in diameter
and 1 m deep, and by gently stirring (1 rpm) the sample
during settling.
6.2.4.2 Sample handling and transport
Long-distance transport and prolonged storage of the
sample should be avoided. Agitation of the sample
during transport and biological activity during storage
may severely influence the settling behaviour. If
possible, perform the measurements on-site at the
WWTP immediately after sampling.
The measurement times and dilution series described in
sections 6.2.2 and 6.2.3 can be considered as a minimum
set of measurements for a batch curve. However, more
frequent measurement times or additional dilutions can
be added depending on the case-specific conditions or
requirements. For example, if the solid-liquid interface
shows a very rapid initial increase then additional
measurements can be recorded during the first 2 min of
sampling. If the appropriate specialised equipment and
experience is available then the solid-liquid interface
may even be tracked automatically (Vanderhasselt and
Vanrolleghem, 2000). Depending on the concentration of
the recycle flow sample, the standard dilution series (100,
80, 60, 50, 40 and 20 %) may not provide sufficient
coverage of the concentration range for hindered settling.
Additional dilutions such as 55 %, 30 % etc. can be
added.
6.2.5 Recent advances in batch settling tests
By measuring a batch curve, the velocity at which the
suspension-liquid interface passes through different
settling regions can be investigated. However, the
drawback of this method is that it only provides
information on the suspension-liquid interface. No
information on the settling behaviour inside the sludge
blanket or the actual build-up of the sludge blanket is
recorded. Nor does it allow the settling velocity in the
246
compression stage to be calculated. More advanced
measurement techniques aiming to provide more detailed
information on the sludge settling behaviour over time
and depth have been presented in dedicated literature.
Examples include detailed spatio-temporal solids
concentration measurements by means of a radioactive
tracer (De Clercq et al., 2005) and velocity measurements
throughout the depth of the sludge blanket with an
ultrasonic transducer (Locatelli et al., 2015). However,
these techniques require specialised equipment and
cannot be routinely performed.
6.3 MEASURING FLOCCULATION STATE
OF ACTIVATED SLUDGE
As can be seen from Figure 6.1, the settling behaviour of
a sludge sample is not only influenced by its
concentration but also by its flocculation state. Hence, the
ability of an SST to act successfully as a clarifier is highly
dependent on the potential of the microorganisms to form
a flocculent biomass which settles and compacts well,
producing a clear effluent (Das et al., 1993). The aim of
the flocculation process is to combine individual flocs
into large and dense flocs that settle rapidly and to
incorporate discrete particles that normally would not
settle alone. If the flocculation process or breakup of
flocs fails during the activated sludge process, a fraction
of the particles is not incorporated into flocs. They
remain in the supernatant of the SST due to lack of
sufficient mass and are carried over the effluent weir,
reducing the effluent quality. Failure of the settling
process can have multiple causes: (i) denitrifying sludge,
(ii) excessively high sludge blankets, (iii) poor
flocculation, or (iv) poor hydrodynamics. In order to take
appropriate remedial actions, it is important to be able to
pinpoint the cause of the failure. Denitrifying sludge and
high sludge blankets can be easily recognised and
corrected (Parker et al., 2000). Distinguishing between
flocculation problems and poor hydrodynamics is more
challenging but can be accomplished by means of a
Dispersed Suspended Solids/Flocculated Suspended
Solids (DSS/FSS) test (Wahlberg et al., 1995)
6.3.1 DSS/FSS test
6.3.1.1 Goal and application
Wahlberg et al. (1995) proposed a procedure that makes
it possible to distinguish between hydraulic and
flocculation problems in a given SST, the so-called
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
DSS/FSS test. Using the DSS and/or FSS test has been
proven to be a useful technique in several studies: it
makes it possible (i) to assess the flocculation and
deflocculation processes in transmission channels (Das et
al., 1993; Parker and Stenquist, 1986; Parker et al.,
1970), (ii) to determine the influence of hydraulic
disturbances in the aeration basin on the effluent nonsettleable sludge particles (Das et al., 1993; Parker et al.,
2000, 1970) and (iii) to determine the benefits of a
flocculation procedure in decreasing effluent suspended
solids in a WWTP (Parker et al., 2000; Wahlberg et al.,
1994).
The DSS/FSS test can be divided into three parts: the
Effluent Suspended Solids (ESS) test, the Dispersed
Suspended Solids (DSS) test, and the Flocculated
Suspended Solids (FSS) test. The ESS test consists of a
simple TSS test to determine the effluent concentration.
The procedures for the DSS and FSS test are provided in
sections 6.3.1.3 and 6.3.1.4.
6.3.1.2 Equipment
The following equipment is needed for the execution of
ESS, DSS and FSS tests:
General
a. Equipment for the TSS test (APHA et al., 2012).
b. A stopwatch.
ESS test
a. An effluent sample (minimum 0.5 L).
DSS test
b. A Kemmerer sampler.
c. A siphon for supernatant sampling.
FSS test
a. A square flocculation jar of at least 2.0 L.
b. Six paddle stirrers.
c. An activated sludge sample (minimum 1.5 L).
6.3.1.3 DSS test
Dispersed Suspended Solids (DSS) are defined as the
concentration of SS remaining in the supernatant after 30
min of settling (Parker et al., 1970). The DSS test thus
quantifies an activated sludge’s state of flocculation at
the moment and location the sample is taken. This is
accomplished by the use of a single container (i.e. a
Kemmerer sampler) for sampling and settling in order to
protect the biological flocs in the sample from any
SETTLING TESTS
secondary flocculation or breakup effects caused by an
intermediate transfer step.
A Kemmerer sampler is a clear 4.2 L container, 105
mm in diameter and 600 mm tall with upper and lower
closures (Figure 6.11). The sample is collected in the
Kemmerer sampler and allowed to settle for 30 min, after
which time the supernatant is sampled using a siphon and
analysed for SS concentration (Wahlberg et al., 1995).
247
• Protocol
1. Immerse the Kemmerer sampler at the desired
location in the SST to grab a sample.
2. Allow this sample to settle for 30 min.
3. After 30 min sample 500 mL of the supernatant (be
careful not to disturb the settled sludge).
4. Analyse the sampled supernatant for SS
concentration.
• Example
Parker et al. (2000) illustrated the use of a DSS test for
SST failure troubleshooting in the case of the Central
Marin Sanitation Agency plant in California. A DSS test
was performed at the plant as high ESS values were being
observed during peak flows. The results of the DSS test
are provided in Table 6.7.
Table 6.7 Measured DSS values at the Central Marin Sanitation Agency
plant in California (Parker et al., 2000).
DSS (mg L-1)
Inlet centre well
10.4
DSS (mg L-1)
Outlet centre well
11.0
DSS (mg L-1)
Effluent weir
3.6
ESS (mg L-1)
Effluent weir
8.5
From these results, it became clear that no
flocculation was occurring in the centre well as the DSS
values at the inlet and outlet of the centre well are similar.
However, the significant reduction in DSS values
between the outlet of the centre well and the effluent weir
showed that flocculation was occurring in the
sedimentation tank. This indicated that the sludge had a
good flocculation tendency but merely lacked proper
conditions for flocculation in the existing centre well,
which was contributing to its small diameter.
Figure 6.11 A Kemmerer sampler with open upper and lower closures
(photo: Royal Eijkelkamp).
The large, settleable flocs settle in the 30 min period,
whereas the dispersed, primary particles not incorporated
in the settling sludge remain in the supernatant. DSS
concentrations have been shown to closely approximate
the ESS concentration for a well-designed and operated
SST (Parker and Stenquist, 1986). Hence, large
deviations between ESS and DSS indicate clarification
problems. DSS tests can be performed with samples at
several locations in the SST (for example: at the SST
inlet, at the outlet of the flocculation well, or near the
effluent weir) to analyse where potential problems (and
for instance flocculation or breakup) are occurring.
Moreover, the DSS results uncovered a significant
hydraulic problem in the clarifiers. The high ESS
concentration compared to the DSS at the effluent weir
signifies the wash out of settleable solids over the
effluent weir of the clarifier.
6.3.1.4 FSS test
Wahlberg et al. (1995) developed a complimentary test
to the DSS test called the Flocculated Suspended Solids
(FSS) test. Whereas the DSS test assesses the state of
flocculation of a sample at a specific location in the SST
at a moment in time, the FSS test quantifies the
flocculation potential of an activated sludge sample by
flocculating the sample under ideal conditions prior to
settling.
248
For this test, an activated sludge sample is collected
in a square flocculation jar (minimum jar volume 2 L).
The sample volume should be at least 1.5 L. The sample
is gently stirred for 30 min at 50 rpm (Figure 6.12) before
it is allowed to settle for 30 min and the concentration in
the supernatant is measured. Flocculation is maximized
by stirring, and settling is performed in an ideal device
(without hydraulic disturbances). Hence, the measured
FSS is considered to be the minimal possible ESS.
Because the FSS concentration is measured under
conditions of maximum flocculation and ideal settling, it
will not change between the aeration basin and the SST.
Therefore the activated sludge sample for the FSS test
can be collected anywhere between the aeration basin and
the SST.
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
• Protocol
1. Collect an activated sludge sample of 1.5 L from the
WWTP.
2. Pour this sample into a square flocculation jar with a
volume of 2 L.
3. Stir the sample for 30 min. at 50 rpm.
4. Stop stirring and allow the sample to settle for
30 min.
5. After 30 min, sample 500 mL of the supernatant
(being careful not to disturb the settled sludge).
6. Analyse the sampled supernatant for suspended
solids concentration.
6.3.1.5 Interpretation of a DSS/FSS test
A well-functioning SST provides proper conditions for
flocculation in order to incorporate small, dispersed
solids that do not have sufficient mass to settle in the SST
into flocs. Failure of the SST with respect to this function
will result in high ESS concentrations. Dispersed
suspended solids exist as a result of three possible
mechanisms; (i) their flocculation is prevented by surface
chemistry reactions (i.e. a biological flocculation
problem), (ii) they are not incorporated into flocs due to
insufficient time for flocculation (i.e. a physical
flocculation problem), or (iii) they have been sheared
from a floc particle due to excessive turbulence (i.e. a
hydraulic problem). A DSS/FSS test makes it possible to
distinguish between these different scenarios in order to
take appropriate remedial actions.
A typical DSS/FSS test consists of four
measurements: the ESS concentration, the DSS
concentration at the inlet of the SST (DSSi), the DSS
concentration at the effluent weir (DSSo) and the FSS
concentration. If a flocculation well is present, then DSSi
should be measured after this structure. Assuming that
the system under study is struggling with high ESS
concentrations, because DSS/FSS tests are typically
performed in the case of clarification failure, four
scenarios can be defined. A DSS/FSS troubleshooting
matrix which shows the cause of the poor performance
under the various testing scenarios is provided in Table
6.8 (Kinnear, 2000).
Table 6.8 DSS/FSS troubleshooting matrix (Kinnear, 2000).
FSS
ESS high and:
Figure 6.12 Experimental setup for the FSS test. An activated sludge
sample is stirred for 30 min in a square flocculation jar (photo: E. Torfs).
DSS
High
Low
High
Biological
flocculation
Not possible
Low
Physical
flocculation
Hydraulics
SETTLING TESTS
249
The different testing scenarios can be interpreted as
follows.
and at the effluent weir. The measured DSS, FSS and
ESS values are shown in Table 6.9.
High DSSi - low FSS
Table 6.9 Measured DSS, FSS and ESS values at the Greeley Water
Pollution Control Facility in Colorado (Brischke et al., 1997).
This scenario indicates poor flocculation in the SST even
though the activated sludge has good flocculating
properties. Either the activated sludge is not receiving
adequate time to flocculate or significant floc breakup is
occurring prior to settlement (for example, by excessive
shear in a conveyance structure). The clarification failure
can be contributing to a flocculation problem of a
physical nature that can be solved by either removing the
cause of breakup or by incorporating an additional
flocculation step prior to settling.
High DSSi - high FSS
As in the high DSSi - low FSS case, these results indicate
poor flocculation in the SST. However, even under the
ideal flocculation circumstances provided by the FSS
test, the clarification cannot be improved. Hence,
additional flocculation will not improve the clarification
and the problem is most likely of a biological nature
resulting in a sludge with poor flocculation properties.
Modifications to the SST will not solve this problem;
attention must be directed upstream of the SST.
Low DSSi - low FSS
The low DSSi suggests that the incoming activated
sludge is in a well-flocculated state. As the DSSi is
already in the same range as the FSS concentration,
further flocculation will not improve the SST
performance and the problem is most likely a hydraulic
one (for example, due to short-circuiting). Comparing the
DSSo and ESS concentration can provide further
confirmation; if the DSSo concentration is significantly
lower than the ESS concentration, then hydraulic
scouring of settleable solids from the sludge blanket is
indicated. To improve the clarification in this case, the
tank’s hydrodynamics need to be investigated by means
of dye tests and/or 2-3D CFD (computational flow
dynamics) modelling.
Low DSSi - high FSS
This outcome is theoretically not possible. Should it
occur it is recommended to repeat the test.
• Example
A DSS/FSS test was used to assess the performance of
the existing clarifiers prior to plant expansion at the
Greeley Water Pollution Control Facility in Colorado
(Brischke et al., 1997; Parker et al., 2000). The DSS test
was performed at two locations i.e. at the inlet to the SST
DSSi (mg L-1)
29.2
DSSo (mg L-1)
22.0
FSS (mg L-1)
8.2
ESS (mg L-1)
25.5
High DSS values both at the inlet and near the
effluent weir indicate that no flocculation is occurring in
the tank. The much lower FSS value signifies that the
sludge has a high potential to flocculate but lacks
appropriate conditions for flocculation in the tank. The
high ESS concentrations can thus be attributed to a
physical flocculation problem that can be solved by
physically modifying the tank in order to provide suitable
flocculation conditions. In this specific case this was
accomplished through modifications of the centre well.
6.3.2 Recommendations
6.3.2.1 Flocculation conditions
Proper execution of an FSS test requires ideal
flocculation conditions. Therefore, it is important to use
a square flocculation jar in order to avoid the formation
of a vortex during mixing. Moreover, make sure that the
sample is completely mixed (i.e. no dead zones at either
the top or bottom of the flocculation jar).
6.3.2.2 Temperature influence
The sample volume for the FSS test is relatively small in
comparison to the volume of the Kemmerer sampler.
Hence, some precautions should be taken to ensure that
the samples do not change drastically in temperature
during the 1 h it takes to conduct the test. For example:
do not perform the test in direct sunlight.
6.3.2.3 Supernatant sampling
Regardless of the specific supernatant sampling
technique, care should be taken not to pull any floating
debris or settled solids into the supernatant sample as this
can severely alter the results. Moreover, as the
concentration in effluent and supernatant is generally
very low, the sampled volume for the TSS test should be
sufficiently large (± 500 mL).
250
6.3.3 Advances in the measurement of the
flocculation state
From the above it becomes clear that ‘good
bioflocculation’ of the activated sludge is a prerequisite
for ‘good sedimentation’ and a good effluent quality.
Hence, what is the definition of a well-flocculated
activated sludge floc? An activated sludge floc is
composed of: (i) a backbone of filamentous organisms,
onto which, (ii) microcolonies (i.e. clusters of microorganisms) can attach and this aggregation of microorganisms is then embedded in a matrix of (iii)
extracellular polymer substances (EPS) (Figure 6.13).
Whenever one of these components is not in balance with
the rest then problems might be encountered.
Figure 6.13 The structural makeup of an activated sludge floc:
microcolonies attach to filamentous bacteria which form the backbone of
the floc, while extracellular polymeric substances constitute the
embedding matrix (Nielsen et al., 2012).
One of the most common settling problems is that of
filamentous bulking where there is a dominance of
filamentous organisms that will make the floc structure
very open and will retain a lot of water. Such filaments
might even entangle with other protruding filaments so
that a network is formed that prevents the sludge from
settling. In contrast, when there are not enough filaments
to form the backbone, so-called pinpoint flocs are
observed. These are small clusters of micro-colonies that
do not settle well. Hence, a good floc that settles well
should be dense (not open) and sufficiently large.
Such characteristics of activated sludge can be
quantified by microscopic image analysis. Changes in the
average floc size, filament length, floc roundness, floc
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
fractal dimension etc. can reveal a lot of information on
the settling behaviour of sludge. While some research
groups have been developing specific image analysis
software to infer this information (Amaral and Ferreira,
2005; Da Motta et al., 2001a, 2001b; Jenneé et al., 2004;
van Dierdonck et al., 2013), also freeware software is
available (such as ImageJ or FIJI (http://fiji.sc/Fiji)) to
perform a basic analysis. A comprehensive overview of
what is currently available in this image analysis domain
is available in dedicated literature (Mesquita et al., 2013).
For microscopic monitoring that is more focused on
revealing the specific microbial communities present,
FISH analysis can be interesting; the reader is referred to
the FISH handbook for biological wastewater treatment
(Nielsen et al., 2009).
As well as image analysis, three additional
bioflocculation-related monitoring tools can be
mentioned that are more focused on the forces that hold
the floc components together. On the one hand, the global
floc strength measurement (Mikkelsen and Keiding,
2002) compares the turbidity of the supernatant before
and after shearing of a sludge sample. Lower turbidity
after shearing indicates better bioflocculation. On the
other hand, relative hydrophobicity and surface charge
can be measured. The relative hydrophobicity is related
to hydrophobic interactions that prove to be important in
keeping the floc aggregated. The value that results from
such an analysis (e.g. the MATH test, Chang and Lee,
1998) which assesses the microbial adhesion tendency to
hydrocarbons) should not be taken as an absolute value
but can be interesting in revealing changes over a certain
operational period. The surface charge is related to
electrostatic forces; with a more neutral activated sludge
surface, the aggregates experience less repulsion
resulting in improved coagulation and flocculation. The
measurement of surface charge is based on a colloid
titration technique (Kawamura and Tanaka, 1966;
Kawamura et al., 1967; Morgan et al., 1990).
6.4 MEASURING THE SETTLING BEHAVIOUR OF
GRANULAR SLUDGE
6.4.1 Goal and application
In recent years new technologies have been developed to
improve the separation of sludge from the treated
effluent. One of these technologies is the use of aerobic
granular sludge. Aerobic granules are spherical biofilms
with a typical shape factor of 0.7 - 0.8 (Beun et al., 2002).
Whereas conventional activated sludge is characterized
SETTLING TESTS
by settling velocities below 5 m h-1 (Vanderhasselt and
Vanrolleghem, 2000), granular sludge settles
significantly faster with settling velocities in the range of
10 up to 100 m h-1 (Bassin et al., 2012; Etterer and
Wilderer, 2001; Winkler et al., 2012, 2011a).
Moreover, granules show a low flocculating tendency
positioning their settling behaviour at the far left side of
Figure 6.1. The granules will thus settle independent even
at higher concentrations (with almost no hindered and
compression regime present) and directly form a compact
sludge bed. For this reason the SVI after 5 minutes will
be approximately the same as the SVI measured after 30
minutes. Typical SVI values for granules are less than 30
ml/g (de Kreuk and van Loosdrecht, 2004; Liu et al.,
2005; Liu and Tay, 2007; Tay et al., 2004).
Granular sludge is developed in Sequencing Batch
Reactors (SBR) as these systems fulfil a number of
specific requirements for the formation of granules: a
feast - famine regime for the selection of appropriate
microorganisms (Beun et al., 1999), short settling times
to ensure retention of granular biomass and wash-out of
flocculent biomass (Qin et al., 2004) and sufficient shear
force to ensure an optimal physical granule integrity (Tay
et al., 2001).
One of the most important parameters to select for
granular sludge is the settling velocity. By applying short
settling times in an SBR, only large biomass aggregates
that settle well are selected, while flocculent sludge is
washed out (Beun et al., 2000). The parameters
determining the settling velocity of particles and in turn
biomass washout are of crucial importance to granular
sludge technology. The balances of forces for the
sedimentation of a spherical particle depend on the
buoyancy, gravity and drag force (Giancoli, 1995). From
this relation, the settling velocity is influenced by the
water viscosity, particle size and shape, and the
difference between the density of the water and the
particles. Hence, the settling velocity of granules is
influenced by the density and size of the particles where
an increase in diameter affects the settling velocity more
severely (Winkler et al., 2012, 2011a). Therefore, this
section presents a method to measure the density and size
of granules as well as a procedure to calculate the
theoretical settling velocity of granules under different
temperature conditions.
6.4.2 Equipment
For granular sludge tests the following equipment is
needed:
251
a. A pycnometer.
b. A microbalance.
c. Sieves or a microscope with an image analyser.
6.4.3 Density measurements
Granule densities between 1,036 - 1,048 kg m-3 have been
reported (Etterer and Wilderer, 2001), which are
comparable to densities of conventional activated sludge
(1,020-1,060 kg m-3) (Andreadakis, 1993; Dammel and
Schroeder, 1991). However, precipitates may form
within the granule core (Lee and Chen, 2015; Mañas et
al., 2012) and significantly increase the granule density
up to 1,300 kg m-3 (Juang et al., 2010; Winkler et al.,
2013).
The specific biomass density can be measured with a
pycnometer (Figure 6.14). A pycnometer is a simple and
inexpensive glass flask, with an exactly calibrated
volume. The pycnometer flask is closed with a glass
stopper that acts as a valve; it contains a small groove,
through which excess water is forced out when closed.
The pycnometer has a known volume V.
Figure 6.14 Picture of a pycnometer (photo: E. Torfs).
• Protocol
1. Measure the weight of the pycnometer (closed with
the glass stopper) in a completely dry state (m0).
2. Fill the pycnometer with water and measure its
weight again (mT). It is very important to dry the
pycnometer carefully before the weight is determined
in order to remove all excess water from the outside
of the pycnometer.
252
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
3. Calculate the mass of water in the pycnometer (mH2O)
as:
mH2O = mT ‒ m0
Eq. 6.11
4. Measure the weight of the granule sample for which
you want to determine the density (ms).
5. Place the granule sample inside the pycnometer and
determine the weight of the pycnometer together with
the inserted sample (m0 + ms).
6. Fill the pycnometer (containing the solids sample)
further with water and weigh its mass again (mTS). As
in Step two, make sure that the pycnometer is dried
carefully before the weight is determined.
7. Calculate the weight of the added water (m’H2O) as:
m'H2O = mTS ‒ mo ‒ ms
Eq. 6.12
8. Determine the volume of added water (V’H2O)
according to:
V'H2 O =
m'H2 O
ρH O
Eq. 6.13
2
9. Calculate the volume of measured solids Vs from the
difference between the volume of water that fills the
empty pycnometer (V) and the previously determined
volume of water (V’H2O).
Vs = V ‒ V'H2 O =
mH2 O ‒ m'H2O
ρH O
Eq. 6.14
2
10. Finally, calculate the density of the granules ρs as:
ρs =
ms
Vs
Table 6.10 Density calculation for a granular sludge sample.
Variable
Mass empty pycnometer (g)
Mass pycnometer and water (g)
Mass water (g)
Mass solids (g)
Mass pycnometer, solids and water (g)
Mass added water (g)
Volume added water (L)
Volume solids (L)
Density solids (g L-1)
Symbol
m0
mT
mH2O
ms
mTS
m’H2O
V’H2O
Vs
ρs
Procedure
Measured
Measured
Calculated
Measured
Measured
Calculated
Calculated
Calculated
Calculated
Value
54.51
153.70
99.19
19.56
154.55
80.48
0.08
0.02
1,010.40
6.4.4 Granular biomass size determination
Although there is no common consensus on the minimum
diameter (Bathe et al., 2005), sieves with a diameter of
0.2 mm have been used to determine the minimum size
of granular biomass (Bin et al., 2011; de Kreuk, 2006; Li
et al., 2009). The largest diameter reported is 16 mm
(Zheng et al., 2006), but typically diameters range
between 0.5-3 mm (de Kreuk and van Loosdrecht, 2004;
Shi et al., 2009; Winkler et al., 2011b). The size
distribution can be measured either by simple sieving
tests or by means of an image analyser.
6.4.4.1 Sieving
Granule size can be determined by means of sieves with
different mesh sizes. The screening can be performed
with sieves with mesh openings of, i.e. 2.0, 1.0, 0.5 and
0.3 mm, making it possible to cover the most common
granule size range.
Eq. 6.15
• Example
A pycnometer with a volume (V) of 0.1 L is used to
determine the density of a granular sludge sample. All the
measurements and calculations are provided in Table
6.10 (the density of water, ρH2 O , at 20 °C is 998 g L-1).
Figure 6.15 Stacked sieves with different mesh openings (photo:
Fieldmaster)
SETTLING TESTS
253
• Protocol
1. Measure the total wet weight of a sample.
2. Mount the sieves vertically one on top of the other in
increasing order of mesh opening (from bottom to
top) so that the coarsest mesh is at the top (Figure
6.15).
3. Pour the granule sample onto the sieves.
4. Wash each sieve successively to allow the granules to
move from one sieve to the next.
5. Filter the liquid (which has trickled through all the
sieves) in order to collect particles smaller than
0.3 mm.
6. Backwash each sieve to retain each granule fraction
in a separate beaker.
7. Determine the wet weight, and if needed the dry
weight (TSS), the ash content and the VSS of each
fraction, which will result in the theoretical
percentage of each class size (Laguna et al., 1999).
6.4.4.2 Image analyser
Alternatively, the granule size can be measured by means
of an image analyser using the averaged projected surface
area of the granules. For this method, a sample is
transferred into a petri dish and placed under a stereo
microscope with a fixed magnification (e.g. 7.5 ×
magnification). Each image analysed is recorded by the
image analyser. The Petri dish needs to be turned
multiple times in order to measure different granules.
Different image analysers are available on the market and
each requires different handling. An example of granule
size distribution data, created during the measuring
procedure, is presented in Figure 6.16.
30
Number of granules
• Protocol
The measured average density and diameter of granules
can be used to calculate the theoretical settling velocity.
For particle Reynolds numbers smaller than or equal to
1, Stokes’ law can be used to calculate the settling
velocity of a particle.
vs =
2
g ρp ‒ ρw dp
⋅
⋅
.
ρw
νw
18
Eq. 6.16
The Reynolds numbers can be calculated with the
following equation:
Re = dp ⋅
vs
νw
Eq. 6.17
Where, vs is the sedimentation velocity of a single
particle (m s-1), dp is the particle diameter (m), ρp is the
density of a particle (kg m-3), ρw is the density of the fluid
(kg m-3), g is the gravitational constant (9.81 m s-2), νw is
the kinematic viscosity of water (m2 s-1), and Rep is the
particle Reynolds number.
The density and viscosity of the medium depend on
the temperature and the solutes present in the water. With
increasing temperature, the viscosity and density of the
water decrease. At high temperature, water molecules are
more mobile than at low temperature, resulting in a
decrease of viscosity by a factor of two between 10 and
40 °C (Podolsky, 2000). A table with density and
viscosity values of water at different temperatures can be
found elsewhere.
• Example
An example of a calculation for the settling velocity of a
particle with a diameter of 0.4 mm and 1,010 kg m-3 at
different temperatures is given in Table 6.11 and plotted
in Figure 6.17. Note that the condition of Reynolds
numbers smaller than or equal to 1 is not met for every
temperature. The implications of this will be discussed
further on in this section.
25
20
15
10
5
0
0.18
6.4.5 Calculating the settling velocity of
granules
0.51
0.84
1.17
1.50
1.83
2.16
3.05
Diameter (mm)
Figure 6.16 Particle size distribution of a granular sludge reactor.
254
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
Table 6.11 Settling velocity of a granule with a particle diameter of 0.4
mm and a particle density of 1,010 kg m-3 at different temperatures.
Table 6.12 Settling velocity of a granule with a particle diameter of 0.4
mm and a particle density of 1,050 kg m-3 at different temperatures.
T
νw
ρw
Re
vs
T
νw
ρw
Re
vs
°C
5
10
15
20
25
30
35
40
m2 s-1 1.5e-06 1.3e-06 1.1e-06 1.0e-06 9.4e-07 8.2e-07 7.4e-07 6.6e-07
kg m-3 1,000 1,000 999 998 997 996 994 992
0.2
0.2
0.2
0.4
0.5
0.7
1.0
1.5
m h-1
2.1
2.5
3.0
3.6
4.2
5.5
6.8
8.7
10
°C
5
10
15
20
25
30
35
40
m2 s-1 1.5e-06 1.3e-06 1.1e-06 1.0e-06 9.4e-07 8.2e-07 7.4e-07 6.6e-07
kg m-3 1,000 1,000 999 998 997 996 994 992
0.8
1.0
1.4
1.7
2.1
2.8
3.6
4.7
m h-1
10.7 12.1 14.0 15.8 17.7 20.9 23.9 27.9
35
Dense granule
Light granule
30
Settling velocity (m h -1)
Settling velocity (m h -1)
8
6
4
25
20
15
10
2
5
0
0
0
5
10
15
20
25
30
35
40
Temperature (˚C)
Figure 6.17 Calculated settling velocities at different temperatures for
granules with a diameter of 0.4 mm and a density of 1,010 kg m-3.
The example in Figure 6.17 shows that the settling
velocity of a small granule with a diameter of 400 μm and
a density of 1,010 kg m-3 varies between 2 m h-1 and 9 m
h-1 for temperatures ranging between 5 – 40 °C (Winkler
et al., 2012). At low temperatures the separation of small
and light granules from flocs (with typical settling
velocities below ± 5 m h-1) can therefore become
troublesome. Earlier research has experimentally proven
that a start-up process at cold temperatures is difficult. In
addition all microbial processes run slower at low
temperatures (Brdjanovic et al., 1997; Kettunen and
Rintala, 1997; Lettinga et al., 2001), hence limiting
granulation at a lower temperature even further. An
increase in density or diameter of the granules will
significantly increase the settling velocity and thus
facilitate the separation.
• Example
Calculated settling velocities for a particle with the same
diameter (0.4 mm) but a higher density (1,050 kg m-3) at
different temperatures are shown in Table 6.12 and
Figure 6.18.
0
5
10
15
20
25
30
35
40
Temperature (˚C)
Figure 6.18 Calculated settling velocities at different temperatures for a
light (1,010 kg m-3) and dense (1,050 kg m-3) granular sludge particle at a
diameter of 0.4 mm.
Larger and denser particles result in particle
Reynolds numbers larger than 1. For these particles the
theoretical settling velocities calculated according to
Stokes’ law will deviate from the true settling velocities.
For most applications, these errors (< 10 %) are
acceptable and Stokes’ law remains a good
approximation. More accurate velocity calculations are
possible but require an advanced calculation method
including an iterative procedure to determine Re.
Illustrations of this approach can be found in Winkler et
al. (2012).
6.4.6 Recommendations
6.4.6.1 Validation of results
The calculated settling velocities (Section 6.4.5) can be
validated experimentally by conducting simple settling
tests with granules harvested from sieves with different
mesh sizes (Section 6.4.4). The different size fractions
can be poured into the reactor column itself or into a
SETTLING TESTS
cylinder and the time is measured until the granule
reaches the bottom of the reactor in order to express its
settling velocity in meters per hour.
6.4.6.2 Application for flocculent sludge
The experimental methods described in sections 6.4.3
and 6.4.4 can also be applied to activated sludge in SSTs
(particularly in the top region of SSTs where discrete
settling is known to occur). However, this type of
application is currently not very common because the
high flocculation potential of sludge causes attention to
be mainly directed towards hindered settling (as this is
the dominant regime at the concentration where activated
sludge enters the SST and has been mainly used to
determine SSTs’ capacity and design). For granular
sludge on the other hand, the dominant settling regime is
discrete thus causing the focus to shift to size and density
measurements.
6.5 MEASURING SETTLING VELOCITY
DISTRIBUTION IN PSTs
6.5.1 Introduction
Primary settling tanks (PSTs) are used as a pre-treatment
in WWTPs. Measurement campaigns conducted since
the early 1970s on urban discharges have clearly shown
that many pollutants occur in particulate form. Moreover,
particles transported in suspension have also emerged as
highly settleable despite their relatively small particle
size (30 to 40 µm). Hence, gravitational separation in
PSTs can serve as a valuable tool to separate coarse,
settleable particles in the raw wastewater prior to further
treatment in the biological reactors.
As concentrations of particulate matter in raw
wastewater are relatively low compared to concentrations
in SSTs, the settling regime in PSTs has a discrete (nonflocculent and flocculent) nature and the settling velocity
depends on the individual properties of particles. Given
the variety of densities, shapes and sizes of suspended
particles in wastewaters (have in mind Stokes’ law), it is
challenging and time-consuming to calculate the settling
velocities of different particles from size and density
measurements. Therefore, this section presents a method
to directly measure the distribution of settling velocities
in representative wastewater samples. This measurement
can be performed using the ViCAs protocol, developed
by Chebbo and Gromaire (2009). The ViCAs (‘Vitesse
255
de Chute en Assainissement’, which is French for
‘settling velocity in sanitation’) is a test to measure the
settling velocity of particles in a column under static
conditions. The test provides insight into the behaviour
of particles present in a wastewater sample in order to
obtain an idea about its composition. As such it can serve
as important information in different application
domains. In primary settling tanks, results from ViCAs
experiments can be used as input to primary clarifier
models (Bachis et al., 2015) or to study the effect of
chemical dosage on the settling velocity distribution in
order to improve chemically enhanced primary treatment
(CEPT) performance. Furthermore, ViCAs experiments
can be applied in the design of combined sewer retention
tanks where knowledge of the settling velocity
distribution can be used to determine an optimal HRT
and corresponding load reduction to the treatment plant.
6.5.2 General principle
A ViCAs test is performed by settling a sample in a
ViCAs column (Figure 6.19).
Pumping
Pumping tube
tube
Valve
¼ valve
turn
¼ turn
Fixing
Fixingscrew
screws
Sedimentation
Sedimentation
column
column
Sample
Samplebox
box
Cups Cups
Figure 6.19 The ViCAs test equipment (photo: Chebbo and Gromaire,
2009).
The ViCAs protocol is based on the principle of
homogeneous suspensions (Figure 6.20). At the
beginning of the measurement, the solids are uniformly
distributed over the whole sedimentation height. Then the
particles are assumed to settle independently of each
256
EXPERIMENTAL METHODS IN WASTEWATER TREATMENT
100
Mass fraction (%) of particles with
settling velocity less than vs (m h-1)
other, without forming aggregates and without diffusion.
The solids, after having settled for a predetermined
period of time, are recovered at the bottom of the
sedimentation column in cups.
90
80
70
60
50
40
30
20
10
0
0.00
0.01
0.10
1.00
10.00
100.00
Settling velocity vs (m h-1)
t=0
M(t)
Figure 6.22 Settling velocity distribution curve f(vs) for a typical
wastewater sample.
Figure 6.20 Principle of the homogeneous suspension.
Their mass is thus recovered from the cups, which
allows the evolution of the cumulative mass M(t) of
settled material as a function of time t to be determined
(Figure 6.21).
1,600
Settled mass of TSS
1,400
1,200
1,000
800
600
400
200
0
0 100
300
500
700
900 1,100
Settling time (min)
1,300 1,500
Figure 6.21 Cumulative evolution of the settled mass as a function of time.
In practice, the cumulative curve of settled mass
consists of n points (7 < n < 12), corresponding to n
samples taken after different settling times.
The measurements of the settled mass as a function
of time make it possible to calculate the settling velocity
distribution f(vs). Figure 6.22 shows an example of a
settling velocity distribution curve for a typical
wastewater sample.
6.5.3 Sampling and sample preservation
The test can be carried out using a composite sample or
by mixing several samples with similar composition (e.g.
harvest 5 bottles of 1 L over an interval
![Tríptico del curso [PDF 450KB]](http://s2.studylib.es/store/data/006258131_1-103c5de49e5ed6e7b549fd7206c4fb9a-300x300.png)
